Nimble's WebSocket API (Application Programming Interface) enables users to subscribe to real-time streaming data and events. Behind this API is a software layer that routes data, arbitrates resources and executes Deep Learning and analytics functions to enable enriched business insights at world-class speed.
The latest online documentation can be found here.
Nimble and the WebSocket API have six key elements: Configuration, Source, Sink, Pipeline, Channel, and Stream.
A Configuration is a collection of Sources, Sinks and Pipeline Elements arranged to execute the desired analytic behavior.
A Source brings data into Nimble.
A Sink pushes data out of Nimble.
A Pipeline is made up of processing elements that perform some analytics task.
Since a Source can be shared between multiple Pipelines, a Channel is an abstraction of both a Source and a Pipeline that can be thought of as "the data brought in by this Source is operated on by this Pipeline".
A Stream is a proxied fMP4 video stream from a Source.
There are three WebSocket topic categories:
In the context of this API, PUB
and SUB
do not have the traditional meaning:
Publishing (PUB
) to a WebSocket does not redistribute that data to other subscribers - only the server (Nimble) sends data to subscribers.
Instead, publishing allows the client to configure an open WebSocket session with the server according to the endpoint's PUB
specification.
On the other hand, an endpoint's SUB
specification defines the messages that clients will receive over the connection from Nimble.
Note: Only after a client subscribes to an endpoint (by establishing the WebSocket connection) can it publish JSON to configure the session.
Notification events for changes to the Configuration.
Specify which events to subscribe to.
The initial subscription state for all events.
An initial set of events to subscribe to.
Additional properties are NOT allowed.
Accepts the following message:
Subscribe or unsubscribe from Configuration events.
The subscription state for all events.
A subscription change request.
Additional properties are NOT allowed.
Enable all events.
Enable only Source events.
{
"source": true
}
Disable Source and enable Pipeline events.
{
"source": false,
"pipeline": true
}
Disable all events (verbose).
{
"source": false,
"pipeline": false,
"sink": false
}
Notification events for changes to the Configuration.
Open a WebSocket connection to receive notification events for changes in the Configuration.
The initial subscription state for all events.
An initial set of events to subscribe to.
Additional properties are NOT allowed.
Accepts the following message:
Global Configuration change notification.
Additional properties are allowed.
Source ID
Short-form Source address.
Long-form Source address.
URL without username and password.
Username for HTTP authentication.
Password for HTTP authentication.
Additional properties are allowed.
Additional properties are allowed.
Stream ID
Additional properties are allowed.
Source ID
Short-form Source address.
Long-form Source address.
URL without username and password.
Username for HTTP authentication.
Password for HTTP authentication.
Additional properties are allowed.
Additional properties are allowed.
Stream ID
Additional properties are allowed.
Source ID
Additional properties are allowed.
Additional properties are allowed.
Source ID
Additional properties are NOT allowed.
Additional properties are allowed.
Sink ID
A list of sources to include. An empty list indicates no filtering is enabled.
Source ID
Additional properties are allowed.
Additional properties are allowed.
Sink ID
A list of sources to include. An empty list indicates no filtering is enabled.
Source ID
Additional properties are allowed.
Additional properties are allowed.
Sink ID
Additional properties are allowed.
Additional properties are allowed.
Sink ID
Additional properties are NOT allowed.
Additional properties are allowed.
Pipeline ID
Sequence of processing stages
Short-form Pipeline element.
Long-form Pipeline element.
The name of the Pipeline element.
The type of hardware device to use to run the Pipeline element.
High-throughput ("a") or low-latency ("s") mode.
A mapping from Source IDs to static metadata JSON.
This enables configuring different static metadata for the Pipeline based on the Source. The static metadata is served for the Channel produced from the Source and Pipeline. The keys of this object are Source IDs and the values are arbitrary JSON objects. Multiple Source IDs may be included in each key by separating them with commas.
If this property is omitted, then an empty JSON object is used by default for unspecified sources. Using an empty string key in this mapping allows overriding the default JSON object.
A Pipeline element configuration used for the input source(s).
Additional properties are allowed.
Additional properties are allowed.
The high-level description of the analytics performed by the pipeline.
IDs of input sources to the pipeline
Source ID
IDs of output sinks from the pipeline
Sink ID
Additional properties are allowed.
Additional properties are allowed.
Pipeline ID
Sequence of processing stages
Short-form Pipeline element.
Long-form Pipeline element.
The name of the Pipeline element.
The type of hardware device to use to run the Pipeline element.
High-throughput ("a") or low-latency ("s") mode.
A mapping from Source IDs to static metadata JSON.
This enables configuring different static metadata for the Pipeline based on the Source. The static metadata is served for the Channel produced from the Source and Pipeline. The keys of this object are Source IDs and the values are arbitrary JSON objects. Multiple Source IDs may be included in each key by separating them with commas.
If this property is omitted, then an empty JSON object is used by default for unspecified sources. Using an empty string key in this mapping allows overriding the default JSON object.
A Pipeline element configuration used for the input source(s).
Additional properties are allowed.
Additional properties are allowed.
The high-level description of the analytics performed by the pipeline.
IDs of input sources to the pipeline
Source ID
IDs of output sinks from the pipeline
Sink ID
Additional properties are allowed.
Additional properties are allowed.
Pipeline ID
Additional properties are allowed.
Additional properties are allowed.
Pipeline ID
Additional properties are NOT allowed.
Live video Source created.
{
"type": "source",
"action": "created",
"value": {
"id": 2,
"type": "video",
"address": "rtsp://localhost:554/axis-media/media.amp",
"rate": 30
}
}
Live video Source updated (address and rate).
{
"type": "source",
"action": "updated",
"value": {
"id": 2,
"type": "video",
"address": "rtsp://localhost:8554/h264major",
"rate": 15
}
}
Live video Source stats updated.
{
"type": "source",
"action": "stats",
"value": {
"status": "running",
"fps": 15,
"latency": 12.85
}
}
Live video Source deleted.
{
"type": "source",
"action": "deleted",
"value": {
"id": 2
}
}
File Sink created.
{
"type": "sink",
"action": "created",
"value": {
"id": 1,
"type": "file",
"address": "./output.mp4",
"filter": []
}
}
File Sink updated (address and filter).
{
"type": "sink",
"action": "updated",
"value": {
"id": 1,
"type": "video",
"address": "./new_output.mp4",
"filter": [
2
]
}
}
File Sink stats updated.
{
"type": "sink",
"action": "stats",
"value": {
"status": "running",
"fps": 14.8,
"latency": 147.2
}
}
File Sink deleted.
{
"type": "sink",
"action": "deleted",
"value": {
"id": 1
}
}
Pipeline created.
{
"type": "pipeline",
"action": "created",
"value": {
"id": 0,
"elements": [
"person-detection-nano:a:0.3:0.3",
"people-tracking-OH:3:3:50:bottom:0.2:3"
],
"usecase": "my-usecase",
"sources": [
2
],
"sinks": [
1
]
}
}
Pipeline updated (usecase and sinks).
{
"type": "pipeline",
"action": "updated",
"value": {
"id": 0,
"elements": [
"person-detection-nano:a:0.3:0.3",
"people-tracking-OH:3:3:50:bottom:0.2:3"
],
"usecase": "new-usecase",
"sources": [
2
],
"sinks": []
}
}
Pipeline stats updated.
{
"type": "pipeline",
"action": "stats",
"value": {
"status": "running",
"fps": 14.9,
"latency": 122.67
}
}
Pipeline deleted.
{
"type": "pipeline",
"action": "deleted",
"value": {
"id": 0
}
}
Notification events for changes to the Exports.
Specify which events to subscribe to.
The initial subscription state for all events.
An initial set of events to subscribe to.
Additional properties are NOT allowed.
Accepts the following message:
Subscribe or unsubscribe from events.
The subscription state for all events.
A subscription change request.
Additional properties are NOT allowed.
Enable all events.
Enable only Channel events.
{
"channel": true
}
Disable Channel and enable Stream events.
{
"channel": false,
"stream": true
}
Disable all events (verbose).
{
"channel": false,
"stream": false
}
Notification events for changes to the Exports.
Open a WebSocket connection to receive notification events for changes to the Exports.
The initial subscription state for all events.
An initial set of events to subscribe to.
Additional properties are NOT allowed.
Accepts the following message:
Exports changed notification
Additional properties are allowed.
Channel ID
Source ID
Pipeline ID
Stream ID
Additional properties are allowed.
Additional properties are allowed.
Channel ID
Additional properties are allowed.
Additional properties are allowed.
Stream ID
Source ID
Additional properties are allowed.
Additional properties are allowed.
Stream ID
Additional properties are allowed.
Stream created.
{
"type": "stream",
"action": "created",
"value": {
"id": 0,
"sources": [
2
]
}
}
Channel created.
{
"type": "channel",
"action": "created",
"value": {
"id": 1,
"source": 2,
"pipeline": 3,
"stream": 0
}
}
Channel deleted.
{
"type": "channel",
"action": "deleted",
"value": {
"id": 1
}
}
Stream deleted.
{
"type": "stream",
"action": "deleted",
"value": {
"id": 0
}
}
Multi-Channel metadata streaming.
Specify which channels to subscribe to.
The initial subscription state for all channels.
An initial set of channels to subscribe to.
Channel ID
Additional properties are allowed.
Name of supported metadata formatter
Additional properties are allowed.
Accepts the following message:
Specify which Channels to subscribe to.
The subscription state for all events.
The Channels to subscribe to.
Channel ID
Enable all Channels.
Enable only Channels 0 and 1.
[
0,
1
]
Disable Channel 1.
[
1
]
Multi-Channel metadata streaming.
Receive JSON metadata from all inferencing Channels.
The initial subscription state for all channels.
An initial set of channels to subscribe to.
Channel ID
Additional properties are allowed.
Name of supported metadata formatter
Additional properties are allowed.
Accepts the following message:
Metadata for a frame of video extracted by the inferencing Channel.
Multi-part metadata message without timestamp.
Channel ID
JSON metadata (based on configured use-case).
Important: Any position or size in the metadata is normalized
from 0
to 1
relative to the dimensions of the video. 0,0
indicates top-left.
For example, if the video dimensions are 1920x1080
pixels, then 0.5,0.2
encodes the position 960,216
in pixels.
This allows the video to be transformed without requiring the metadata to be parsed and transformed too.
Metadata for the Fire Detection use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Metadata for the Line Crossing use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
The IDs of lines that are being violated by the instance.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Mask Detection use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the People Counting use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Physical Distancing use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Pose Detection use-case
The classification name of the pose.
A percentage-based confidence of the classification.
The identified instances of the pose.
The identified part of the pose.
A percentage-based confidence of the classification.
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Metadata for the PPE Compliance use-case
The classification name.
A percentage-based confidence of the classification.
The identified people.
An instance which can be blurred using its bounding box.
PPE gear
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
An instance which can be blurred using its bounding box.
PPE gear
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Tracing use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
Unique tracking ID of the instance.
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
A 2-dimensional point (x,y) using normalized coordinates (0 to 1, inclusive).
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Zone Violation use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
A list of zones (by ID) that this instance is violating.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Optional Jpeg image
Multi-part metadata message with timestamp.
This message format is used when the stream
property of the channel is non-null.
Channel ID
JSON metadata (based on configured use-case).
Important: Any position or size in the metadata is normalized
from 0
to 1
relative to the dimensions of the video. 0,0
indicates top-left.
For example, if the video dimensions are 1920x1080
pixels, then 0.5,0.2
encodes the position 960,216
in pixels.
This allows the video to be transformed without requiring the metadata to be parsed and transformed too.
Metadata for the Fire Detection use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Metadata for the Line Crossing use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
The IDs of lines that are being violated by the instance.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Mask Detection use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the People Counting use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Physical Distancing use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Pose Detection use-case
The classification name of the pose.
A percentage-based confidence of the classification.
The identified instances of the pose.
The identified part of the pose.
A percentage-based confidence of the classification.
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Metadata for the PPE Compliance use-case
The classification name.
A percentage-based confidence of the classification.
The identified people.
An instance which can be blurred using its bounding box.
PPE gear
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
An instance which can be blurred using its bounding box.
PPE gear
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Tracing use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
Unique tracking ID of the instance.
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
A 2-dimensional point (x,y) using normalized coordinates (0 to 1, inclusive).
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Zone Violation use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
A list of zones (by ID) that this instance is violating.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Media presentation timestamp (PTS)
Optional Jpeg image
[
0,
{
"Labels": [
{
"Name": "string",
"Confidence": 1,
"Instances": [
{
"BoundingBox": [
1,
1,
1,
1
],
"Confidence": 1
}
]
}
]
}
]
Multi-Channel event notification.
Specify which channels to subscribe to.
The initial subscription state for all channels.
An initial set of channels to subscribe to.
Channel ID
Additional properties are allowed.
Accepts the following message:
Specify which Channels to subscribe to.
The subscription state for all events.
The Channels to subscribe to.
Channel ID
Enable all Channels.
Enable only Channels 0 and 1.
[
0,
1
]
Disable Channel 1.
[
1
]
Multi-Channel event notification.
Receive notification events for updates to Channels.
The initial subscription state for all channels.
An initial set of channels to subscribe to.
Channel ID
Additional properties are allowed.
Accepts the following message:
Channel statistics changed notification.
Channel ID
The processing frame rate of the channel
Additional properties are allowed.
The processing frame rate of the channel (measured in frames per second)
The processing latency of the channel
Additional properties are allowed.
The processing latency of the channel (measured in milliseconds)
FPS changed event (for Channel 0).
{
"id": 0,
"type": "fps",
"value": 29.2
}
Latency changed event (for Channel 1).
{
"id": 1,
"type": "latency",
"value": 134.1
}
Channel metadata streaming.
Receive JSON metadata from an inferencing Channel.
The ID of the Channel to connect to.
Name of supported metadata formatter
Additional properties are allowed.
Accepts the following message:
Metadata for a frame of video extracted by the inferencing Channel.
Multi-part metadata message without timestamp.
JSON metadata (based on configured use-case).
Important: Any position or size in the metadata is normalized
from 0
to 1
relative to the dimensions of the video. 0,0
indicates top-left.
For example, if the video dimensions are 1920x1080
pixels, then 0.5,0.2
encodes the position 960,216
in pixels.
This allows the video to be transformed without requiring the metadata to be parsed and transformed too.
Metadata for the Fire Detection use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Metadata for the Line Crossing use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
The IDs of lines that are being violated by the instance.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Mask Detection use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the People Counting use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Physical Distancing use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Pose Detection use-case
The classification name of the pose.
A percentage-based confidence of the classification.
The identified instances of the pose.
The identified part of the pose.
A percentage-based confidence of the classification.
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Metadata for the PPE Compliance use-case
The classification name.
A percentage-based confidence of the classification.
The identified people.
An instance which can be blurred using its bounding box.
PPE gear
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
An instance which can be blurred using its bounding box.
PPE gear
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Tracing use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
Unique tracking ID of the instance.
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
A 2-dimensional point (x,y) using normalized coordinates (0 to 1, inclusive).
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Zone Violation use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
A list of zones (by ID) that this instance is violating.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Optional Jpeg image
Multi-part metadata message with timestamp.
This message format is used when the stream
property of the channel is non-null.
JSON metadata (based on configured use-case).
Important: Any position or size in the metadata is normalized
from 0
to 1
relative to the dimensions of the video. 0,0
indicates top-left.
For example, if the video dimensions are 1920x1080
pixels, then 0.5,0.2
encodes the position 960,216
in pixels.
This allows the video to be transformed without requiring the metadata to be parsed and transformed too.
Metadata for the Fire Detection use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Metadata for the Line Crossing use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
The IDs of lines that are being violated by the instance.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Mask Detection use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the People Counting use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Physical Distancing use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Pose Detection use-case
The classification name of the pose.
A percentage-based confidence of the classification.
The identified instances of the pose.
The identified part of the pose.
A percentage-based confidence of the classification.
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Metadata for the PPE Compliance use-case
The classification name.
A percentage-based confidence of the classification.
The identified people.
An instance which can be blurred using its bounding box.
PPE gear
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
An instance which can be blurred using its bounding box.
PPE gear
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Tracing use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
Unique tracking ID of the instance.
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
A 2-dimensional point (x,y) using normalized coordinates (0 to 1, inclusive).
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Zone Violation use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
A list of zones (by ID) that this instance is violating.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Media presentation timestamp (PTS)
Optional Jpeg image
[
{
"Labels": [
{
"Name": "string",
"Confidence": 1,
"Instances": [
{
"BoundingBox": [
1,
1,
1,
1
],
"Confidence": 1
}
]
}
]
}
]
Channel event notification.
Specify which events to subscribe to.
The ID of the Channel to connect to.
The initial subscription state for all events.
An initial set of events to subscribe to.
Additional properties are NOT allowed.
Accepts the following message:
Subscribe or unsubscribe from Channel events.
The subscription state for all events.
A subscription change request.
Additional properties are NOT allowed.
Enable all events.
Enable FPS events.
{
"fps": true
}
Disable FPS and enable latency events.
{
"fps": false,
"latency": true
}
Disable all events (verbose).
{
"fps": false,
"latency": false
}
Channel event notification.
Receive notification events for updates to specific Channel.
The ID of the Channel to connect to.
The initial subscription state for all events.
An initial set of events to subscribe to.
Additional properties are NOT allowed.
Accepts the following message:
Channel statistic changed notification.
The processing frame rate of the channel
Additional properties are allowed.
The processing frame rate of the channel (measured in frames per second)
The processing latency of the channel
Additional properties are allowed.
The processing latency of the channel (measured in milliseconds)
FPS changed event.
{
"type": "fps",
"value": 29.2
}
Latency changed event.
{
"type": "latency",
"value": 134.1
}
Video streaming.
Receive a binary stream of fragmented MP4 video.
The ID of the Stream to connect to.
Accepts one of the following messages:
Stream status indication.
Indicates the video has reached end-of-stream and is restarting.
Reset status.
reset
string
Notification events for updates to the Statistics.
Specify which events to subscribe to.
The initial subscription state for all events.
An initial set of events to subscribe to.
Additional properties are NOT allowed.
Accepts the following message:
Subscribe or unsubscribe from events.
The subscription state for all events.
A subscription change request.
Additional properties are NOT allowed.
Enable all events.
Enable only CPU Load events.
{
"cpu": true
}
Disable CPU Load and enable Memory events.
{
"cpu": false,
"memory": true
}
Disable all events (verbose).
{
"cpu": false,
"memory": false
}
Notification events for updates to the Statistics.
Receive notification events for updates to Statistics.
The initial subscription state for all events.
An initial set of events to subscribe to.
Additional properties are NOT allowed.
Accepts the following message:
Global statistics changed notification.
The uptime (in seconds) since Nimble started
Additional properties are allowed.
The uptime (in seconds) since Nimble started
The CPU load (measured in CPU threads utilized)
Additional properties are allowed.
CPU utilization (measured in CPU threads)
The amount of memory used (measured in bytes)
Additional properties are allowed.
Amount of virtual and physical memory used (measured in bytes)
Amount of virtual memory used (measured in bytes)
Amount of resident physical memory used (measured in bytes)
Additional properties are allowed.
Uptime (in seconds).
{
"type": "uptime",
"value": 29.763
}
One CPU thread utilized.
{
"type": "cpu",
"value": 1
}
Half of a CPU thread utilized.
{
"type": "cpu",
"value": 0.5
}
2.5GiB virtual memory and 800MiB physical RAM used.
{
"type": "memory",
"value": {
"virtual": 2684354560,
"physical": 838860800
}
}
Subscribe or unsubscribe from Configuration events.
The subscription state for all events.
A subscription change request.
Additional properties are NOT allowed.
Global Configuration change notification.
Additional properties are allowed.
Source ID
Short-form Source address.
Long-form Source address.
URL without username and password.
Username for HTTP authentication.
Password for HTTP authentication.
Additional properties are allowed.
Additional properties are allowed.
Stream ID
Additional properties are allowed.
Source ID
Short-form Source address.
Long-form Source address.
URL without username and password.
Username for HTTP authentication.
Password for HTTP authentication.
Additional properties are allowed.
Additional properties are allowed.
Stream ID
Additional properties are allowed.
Source ID
Additional properties are allowed.
Additional properties are allowed.
Source ID
Additional properties are NOT allowed.
Additional properties are allowed.
Sink ID
A list of sources to include. An empty list indicates no filtering is enabled.
Source ID
Additional properties are allowed.
Additional properties are allowed.
Sink ID
A list of sources to include. An empty list indicates no filtering is enabled.
Source ID
Additional properties are allowed.
Additional properties are allowed.
Sink ID
Additional properties are allowed.
Additional properties are allowed.
Sink ID
Additional properties are NOT allowed.
Additional properties are allowed.
Pipeline ID
Sequence of processing stages
Short-form Pipeline element.
Long-form Pipeline element.
The name of the Pipeline element.
The type of hardware device to use to run the Pipeline element.
High-throughput ("a") or low-latency ("s") mode.
A mapping from Source IDs to static metadata JSON.
This enables configuring different static metadata for the Pipeline based on the Source. The static metadata is served for the Channel produced from the Source and Pipeline. The keys of this object are Source IDs and the values are arbitrary JSON objects. Multiple Source IDs may be included in each key by separating them with commas.
If this property is omitted, then an empty JSON object is used by default for unspecified sources. Using an empty string key in this mapping allows overriding the default JSON object.
A Pipeline element configuration used for the input source(s).
Additional properties are allowed.
Additional properties are allowed.
The high-level description of the analytics performed by the pipeline.
IDs of input sources to the pipeline
Source ID
IDs of output sinks from the pipeline
Sink ID
Additional properties are allowed.
Additional properties are allowed.
Pipeline ID
Sequence of processing stages
Short-form Pipeline element.
Long-form Pipeline element.
The name of the Pipeline element.
The type of hardware device to use to run the Pipeline element.
High-throughput ("a") or low-latency ("s") mode.
A mapping from Source IDs to static metadata JSON.
This enables configuring different static metadata for the Pipeline based on the Source. The static metadata is served for the Channel produced from the Source and Pipeline. The keys of this object are Source IDs and the values are arbitrary JSON objects. Multiple Source IDs may be included in each key by separating them with commas.
If this property is omitted, then an empty JSON object is used by default for unspecified sources. Using an empty string key in this mapping allows overriding the default JSON object.
A Pipeline element configuration used for the input source(s).
Additional properties are allowed.
Additional properties are allowed.
The high-level description of the analytics performed by the pipeline.
IDs of input sources to the pipeline
Source ID
IDs of output sinks from the pipeline
Sink ID
Additional properties are allowed.
Additional properties are allowed.
Pipeline ID
Additional properties are allowed.
Additional properties are allowed.
Pipeline ID
Additional properties are NOT allowed.
Subscribe or unsubscribe from events.
The subscription state for all events.
A subscription change request.
Additional properties are NOT allowed.
Exports changed notification
Additional properties are allowed.
Channel ID
Source ID
Pipeline ID
Stream ID
Additional properties are allowed.
Additional properties are allowed.
Channel ID
Additional properties are allowed.
Additional properties are allowed.
Stream ID
Source ID
Additional properties are allowed.
Additional properties are allowed.
Stream ID
Additional properties are allowed.
Specify which Channels to subscribe to.
The subscription state for all events.
The Channels to subscribe to.
Channel ID
Subscribe or unsubscribe from Channel events.
The subscription state for all events.
A subscription change request.
Additional properties are NOT allowed.
Channel statistic changed notification.
The processing frame rate of the channel
Additional properties are allowed.
The processing frame rate of the channel (measured in frames per second)
The processing latency of the channel
Additional properties are allowed.
The processing latency of the channel (measured in milliseconds)
Channel statistics changed notification.
Channel ID
The processing frame rate of the channel
Additional properties are allowed.
The processing frame rate of the channel (measured in frames per second)
The processing latency of the channel
Additional properties are allowed.
The processing latency of the channel (measured in milliseconds)
Metadata for a frame of video extracted by the inferencing Channel.
Multi-part metadata message without timestamp.
JSON metadata (based on configured use-case).
Important: Any position or size in the metadata is normalized
from 0
to 1
relative to the dimensions of the video. 0,0
indicates top-left.
For example, if the video dimensions are 1920x1080
pixels, then 0.5,0.2
encodes the position 960,216
in pixels.
This allows the video to be transformed without requiring the metadata to be parsed and transformed too.
Metadata for the Fire Detection use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Metadata for the Line Crossing use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
The IDs of lines that are being violated by the instance.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Mask Detection use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the People Counting use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Physical Distancing use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Pose Detection use-case
The classification name of the pose.
A percentage-based confidence of the classification.
The identified instances of the pose.
The identified part of the pose.
A percentage-based confidence of the classification.
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Metadata for the PPE Compliance use-case
The classification name.
A percentage-based confidence of the classification.
The identified people.
An instance which can be blurred using its bounding box.
PPE gear
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
An instance which can be blurred using its bounding box.
PPE gear
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Tracing use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
Unique tracking ID of the instance.
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
A 2-dimensional point (x,y) using normalized coordinates (0 to 1, inclusive).
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Zone Violation use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
A list of zones (by ID) that this instance is violating.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Optional Jpeg image
Multi-part metadata message with timestamp.
This message format is used when the stream
property of the channel is non-null.
JSON metadata (based on configured use-case).
Important: Any position or size in the metadata is normalized
from 0
to 1
relative to the dimensions of the video. 0,0
indicates top-left.
For example, if the video dimensions are 1920x1080
pixels, then 0.5,0.2
encodes the position 960,216
in pixels.
This allows the video to be transformed without requiring the metadata to be parsed and transformed too.
Metadata for the Fire Detection use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Metadata for the Line Crossing use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
The IDs of lines that are being violated by the instance.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Mask Detection use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the People Counting use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Physical Distancing use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Pose Detection use-case
The classification name of the pose.
A percentage-based confidence of the classification.
The identified instances of the pose.
The identified part of the pose.
A percentage-based confidence of the classification.
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Metadata for the PPE Compliance use-case
The classification name.
A percentage-based confidence of the classification.
The identified people.
An instance which can be blurred using its bounding box.
PPE gear
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
An instance which can be blurred using its bounding box.
PPE gear
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Tracing use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
Unique tracking ID of the instance.
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
A 2-dimensional point (x,y) using normalized coordinates (0 to 1, inclusive).
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Zone Violation use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
A list of zones (by ID) that this instance is violating.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Media presentation timestamp (PTS)
Optional Jpeg image
Metadata for a frame of video extracted by the inferencing Channel.
Multi-part metadata message without timestamp.
Channel ID
JSON metadata (based on configured use-case).
Important: Any position or size in the metadata is normalized
from 0
to 1
relative to the dimensions of the video. 0,0
indicates top-left.
For example, if the video dimensions are 1920x1080
pixels, then 0.5,0.2
encodes the position 960,216
in pixels.
This allows the video to be transformed without requiring the metadata to be parsed and transformed too.
Metadata for the Fire Detection use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Metadata for the Line Crossing use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
The IDs of lines that are being violated by the instance.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Mask Detection use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the People Counting use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Physical Distancing use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Pose Detection use-case
The classification name of the pose.
A percentage-based confidence of the classification.
The identified instances of the pose.
The identified part of the pose.
A percentage-based confidence of the classification.
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Metadata for the PPE Compliance use-case
The classification name.
A percentage-based confidence of the classification.
The identified people.
An instance which can be blurred using its bounding box.
PPE gear
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
An instance which can be blurred using its bounding box.
PPE gear
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Tracing use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
Unique tracking ID of the instance.
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
A 2-dimensional point (x,y) using normalized coordinates (0 to 1, inclusive).
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Zone Violation use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
A list of zones (by ID) that this instance is violating.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Optional Jpeg image
Multi-part metadata message with timestamp.
This message format is used when the stream
property of the channel is non-null.
Channel ID
JSON metadata (based on configured use-case).
Important: Any position or size in the metadata is normalized
from 0
to 1
relative to the dimensions of the video. 0,0
indicates top-left.
For example, if the video dimensions are 1920x1080
pixels, then 0.5,0.2
encodes the position 960,216
in pixels.
This allows the video to be transformed without requiring the metadata to be parsed and transformed too.
Metadata for the Fire Detection use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Metadata for the Line Crossing use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
The IDs of lines that are being violated by the instance.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Mask Detection use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the People Counting use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Physical Distancing use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Pose Detection use-case
The classification name of the pose.
A percentage-based confidence of the classification.
The identified instances of the pose.
The identified part of the pose.
A percentage-based confidence of the classification.
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Metadata for the PPE Compliance use-case
The classification name.
A percentage-based confidence of the classification.
The identified people.
An instance which can be blurred using its bounding box.
PPE gear
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
An instance which can be blurred using its bounding box.
PPE gear
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Tracing use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
Unique tracking ID of the instance.
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
A 2-dimensional point (x,y) using normalized coordinates (0 to 1, inclusive).
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Zone Violation use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
A list of zones (by ID) that this instance is violating.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Media presentation timestamp (PTS)
Optional Jpeg image
Stream status indication.
Indicates the video has reached end-of-stream and is restarting.
Subscribe or unsubscribe from events.
The subscription state for all events.
A subscription change request.
Additional properties are NOT allowed.
Global statistics changed notification.
The uptime (in seconds) since Nimble started
Additional properties are allowed.
The uptime (in seconds) since Nimble started
The CPU load (measured in CPU threads utilized)
Additional properties are allowed.
CPU utilization (measured in CPU threads)
The amount of memory used (measured in bytes)
Additional properties are allowed.
Amount of virtual and physical memory used (measured in bytes)
Amount of virtual memory used (measured in bytes)
Amount of resident physical memory used (measured in bytes)
Additional properties are allowed.
Additional properties are allowed.
Additional properties are allowed.
Additional properties are allowed.
Source ID
Short-form Source address.
Long-form Source address.
URL without username and password.
Username for HTTP authentication.
Password for HTTP authentication.
Additional properties are allowed.
Additional properties are allowed.
Stream ID
Additional properties are allowed.
Source ID
Short-form Source address.
Long-form Source address.
URL without username and password.
Username for HTTP authentication.
Password for HTTP authentication.
Additional properties are allowed.
Additional properties are allowed.
Stream ID
Additional properties are allowed.
Source ID
Additional properties are allowed.
Additional properties are allowed.
Source ID
Additional properties are NOT allowed.
Additional properties are allowed.
Additional properties are allowed.
Sink ID
A list of sources to include. An empty list indicates no filtering is enabled.
Source ID
Additional properties are allowed.
Additional properties are allowed.
Sink ID
A list of sources to include. An empty list indicates no filtering is enabled.
Source ID
Additional properties are allowed.
Additional properties are allowed.
Sink ID
Additional properties are allowed.
Additional properties are allowed.
Sink ID
Additional properties are NOT allowed.
Additional properties are allowed.
Additional properties are allowed.
Pipeline ID
Sequence of processing stages
Short-form Pipeline element.
Long-form Pipeline element.
The name of the Pipeline element.
The type of hardware device to use to run the Pipeline element.
High-throughput ("a") or low-latency ("s") mode.
A mapping from Source IDs to static metadata JSON.
This enables configuring different static metadata for the Pipeline based on the Source. The static metadata is served for the Channel produced from the Source and Pipeline. The keys of this object are Source IDs and the values are arbitrary JSON objects. Multiple Source IDs may be included in each key by separating them with commas.
If this property is omitted, then an empty JSON object is used by default for unspecified sources. Using an empty string key in this mapping allows overriding the default JSON object.
A Pipeline element configuration used for the input source(s).
Additional properties are allowed.
Additional properties are allowed.
The high-level description of the analytics performed by the pipeline.
IDs of input sources to the pipeline
Source ID
IDs of output sinks from the pipeline
Sink ID
Additional properties are allowed.
Additional properties are allowed.
Pipeline ID
Sequence of processing stages
Short-form Pipeline element.
Long-form Pipeline element.
The name of the Pipeline element.
The type of hardware device to use to run the Pipeline element.
High-throughput ("a") or low-latency ("s") mode.
A mapping from Source IDs to static metadata JSON.
This enables configuring different static metadata for the Pipeline based on the Source. The static metadata is served for the Channel produced from the Source and Pipeline. The keys of this object are Source IDs and the values are arbitrary JSON objects. Multiple Source IDs may be included in each key by separating them with commas.
If this property is omitted, then an empty JSON object is used by default for unspecified sources. Using an empty string key in this mapping allows overriding the default JSON object.
A Pipeline element configuration used for the input source(s).
Additional properties are allowed.
Additional properties are allowed.
The high-level description of the analytics performed by the pipeline.
IDs of input sources to the pipeline
Source ID
IDs of output sinks from the pipeline
Sink ID
Additional properties are allowed.
Additional properties are allowed.
Pipeline ID
Additional properties are allowed.
Additional properties are allowed.
Pipeline ID
Additional properties are NOT allowed.
Additional properties are allowed.
Additional properties are allowed.
Additional properties are allowed.
Channel ID
Source ID
Pipeline ID
Stream ID
Additional properties are allowed.
Additional properties are allowed.
Channel ID
Additional properties are allowed.
Additional properties are allowed.
Additional properties are allowed.
Stream ID
Source ID
Additional properties are allowed.
Additional properties are allowed.
Stream ID
Additional properties are allowed.
The initial subscription state for all channels.
An initial set of channels to subscribe to.
Channel ID
Additional properties are allowed.
Name of supported metadata formatter
Additional properties are allowed.
The processing frame rate of the channel
Additional properties are allowed.
The processing frame rate of the channel (measured in frames per second)
The processing latency of the channel
Additional properties are allowed.
The processing latency of the channel (measured in milliseconds)
Additional properties are allowed.
The processing frame rate of the channel
Additional properties are allowed.
The processing frame rate of the channel (measured in frames per second)
The processing latency of the channel
Additional properties are allowed.
The processing latency of the channel (measured in milliseconds)
Media presentation timestamp (PTS)
Optional Jpeg image
Multi-part metadata message without timestamp.
JSON metadata (based on configured use-case).
Important: Any position or size in the metadata is normalized
from 0
to 1
relative to the dimensions of the video. 0,0
indicates top-left.
For example, if the video dimensions are 1920x1080
pixels, then 0.5,0.2
encodes the position 960,216
in pixels.
This allows the video to be transformed without requiring the metadata to be parsed and transformed too.
Metadata for the Fire Detection use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Metadata for the Line Crossing use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
The IDs of lines that are being violated by the instance.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Mask Detection use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the People Counting use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Physical Distancing use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Pose Detection use-case
The classification name of the pose.
A percentage-based confidence of the classification.
The identified instances of the pose.
The identified part of the pose.
A percentage-based confidence of the classification.
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Metadata for the PPE Compliance use-case
The classification name.
A percentage-based confidence of the classification.
The identified people.
An instance which can be blurred using its bounding box.
PPE gear
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
An instance which can be blurred using its bounding box.
PPE gear
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Tracing use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
Unique tracking ID of the instance.
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
A 2-dimensional point (x,y) using normalized coordinates (0 to 1, inclusive).
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Zone Violation use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
A list of zones (by ID) that this instance is violating.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Optional Jpeg image
Multi-part metadata message with timestamp.
This message format is used when the stream
property of the channel is non-null.
JSON metadata (based on configured use-case).
Important: Any position or size in the metadata is normalized
from 0
to 1
relative to the dimensions of the video. 0,0
indicates top-left.
For example, if the video dimensions are 1920x1080
pixels, then 0.5,0.2
encodes the position 960,216
in pixels.
This allows the video to be transformed without requiring the metadata to be parsed and transformed too.
Metadata for the Fire Detection use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Metadata for the Line Crossing use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
The IDs of lines that are being violated by the instance.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Mask Detection use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the People Counting use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Physical Distancing use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Pose Detection use-case
The classification name of the pose.
A percentage-based confidence of the classification.
The identified instances of the pose.
The identified part of the pose.
A percentage-based confidence of the classification.
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Metadata for the PPE Compliance use-case
The classification name.
A percentage-based confidence of the classification.
The identified people.
An instance which can be blurred using its bounding box.
PPE gear
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
An instance which can be blurred using its bounding box.
PPE gear
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Tracing use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
Unique tracking ID of the instance.
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
A 2-dimensional point (x,y) using normalized coordinates (0 to 1, inclusive).
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Zone Violation use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
A list of zones (by ID) that this instance is violating.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Media presentation timestamp (PTS)
Optional Jpeg image
Multi-part metadata message without timestamp.
Channel ID
JSON metadata (based on configured use-case).
Important: Any position or size in the metadata is normalized
from 0
to 1
relative to the dimensions of the video. 0,0
indicates top-left.
For example, if the video dimensions are 1920x1080
pixels, then 0.5,0.2
encodes the position 960,216
in pixels.
This allows the video to be transformed without requiring the metadata to be parsed and transformed too.
Metadata for the Fire Detection use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Metadata for the Line Crossing use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
The IDs of lines that are being violated by the instance.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Mask Detection use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the People Counting use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Physical Distancing use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Pose Detection use-case
The classification name of the pose.
A percentage-based confidence of the classification.
The identified instances of the pose.
The identified part of the pose.
A percentage-based confidence of the classification.
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Metadata for the PPE Compliance use-case
The classification name.
A percentage-based confidence of the classification.
The identified people.
An instance which can be blurred using its bounding box.
PPE gear
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
An instance which can be blurred using its bounding box.
PPE gear
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Tracing use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
Unique tracking ID of the instance.
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
A 2-dimensional point (x,y) using normalized coordinates (0 to 1, inclusive).
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Zone Violation use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
A list of zones (by ID) that this instance is violating.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Optional Jpeg image
Multi-part metadata message with timestamp.
This message format is used when the stream
property of the channel is non-null.
Channel ID
JSON metadata (based on configured use-case).
Important: Any position or size in the metadata is normalized
from 0
to 1
relative to the dimensions of the video. 0,0
indicates top-left.
For example, if the video dimensions are 1920x1080
pixels, then 0.5,0.2
encodes the position 960,216
in pixels.
This allows the video to be transformed without requiring the metadata to be parsed and transformed too.
Metadata for the Fire Detection use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Metadata for the Line Crossing use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
The IDs of lines that are being violated by the instance.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Mask Detection use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the People Counting use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Physical Distancing use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Pose Detection use-case
The classification name of the pose.
A percentage-based confidence of the classification.
The identified instances of the pose.
The identified part of the pose.
A percentage-based confidence of the classification.
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Metadata for the PPE Compliance use-case
The classification name.
A percentage-based confidence of the classification.
The identified people.
An instance which can be blurred using its bounding box.
PPE gear
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
An instance which can be blurred using its bounding box.
PPE gear
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Tracing use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
Unique tracking ID of the instance.
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
A 2-dimensional point (x,y) using normalized coordinates (0 to 1, inclusive).
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Metadata for the Zone Violation use-case
The classification name.
A percentage-based confidence of the classification.
The identified instances of the object.
An instance which can be blurred using its bounding box.
Instance
A rectangular bounds (x,y,w,h) using normalized coordinates (0 to 1, inclusive).
A percentage-based confidence of the classification.
A list of zones (by ID) that this instance is violating.
Additional properties are NOT allowed.
Additional properties are NOT allowed.
Media presentation timestamp (PTS)
Optional Jpeg image
Indicates the video has reached end-of-stream and is restarting.
Additional properties are allowed.
The uptime (in seconds) since Nimble started
Additional properties are allowed.
The uptime (in seconds) since Nimble started
The CPU load (measured in CPU threads utilized)
Additional properties are allowed.
CPU utilization (measured in CPU threads)
The amount of memory used (measured in bytes)
Additional properties are allowed.
Amount of virtual and physical memory used (measured in bytes)
Amount of virtual memory used (measured in bytes)
Amount of resident physical memory used (measured in bytes)
Additional properties are allowed.