Configuring the Human pose detector

To configure the Human pose detector, do the following:

Go to the Detectors tab.
Below the required camera, click Create… → Category: Poses → Human pose detector.

By default, the detector is enabled and set to detect poses.

If necessary, you can change the detector parameters. The list of parameters is given in the table:

Parameter	Value	Description
Object features
Record objects tracking	Yes	The metadata of the video stream is recorded to the database by default. To disable the parameter, select the No value
Record objects tracking	No
Video stream	Main stream	If the camera supports multistreaming, select the stream for which detection is needed. Selecting a low-quality video stream allows reducing the load on the server.
Video stream	Second stream
Other
Enable	Yes	The detector is enabled by default. To disable, select the No value
Enable	No
Name	Human pose detector	Enter the detector name or leave the default name
Decoder mode	Auto	Select a processing resource for decoding video streams. When you select a GPU, a stand-alone graphics card takes priority (when decoding with NVIDIA NVDEC chips). If there is no appropriate GPU, the decoding will use the Intel Quick Sync Video technology. Otherwise, CPU resources are used for decoding
	CPU
	GPU
	HuaweiNPU
Number of frames processed per second	3	Specify the number of frames that the detector will process per second. The value must be in the range [0.016; 100] Attention! With static persons in the scene, set the FPS to no less than 2. With moving persons in the scene, the FPS must be at least 4. The higher the FPS value, the higher the accuracy of pose detection, but the load on the selected processor is higher as well. For FPS equal to 1, the accuracy is no less than 70%. This parameter varies depending on the object speed of movement. To solve typical tasks, an FPS value from 3 to 20 is sufficient. Examples: pose detection for moderately moving objects (without sudden movements)—FPS 3; pose detection for moving objects—FPS 12.
Type	Human pose detector	Name of the detector type (non-editable field)
Advanced settings If detection of small objects or objects in areas far away from the camera is ineffective, you can use the scanning mode (see Configuring the scanning mode). This mode doesn’t provide absolute detection accuracy but can improve detection performance
Neural network file		Specify a neural network file. The standard neural networks for different processor types are located in the C:\Program Files\Common Files\AxxonSoft\DetectorPack\NeuroSDK directory. You don't need to select the standard neural networks in this field, the system will automatically select the required one. If you use a custom neural network, enter a path to the file. Attention! You cannot specify the network file in Windows OS. You must place the neural network file locally, that is, on the same server where you install Axxon One. For correct neural network operation on Linux OS, place the corresponding file locally in the /opt/AxxonSoft/DetectorPack/NeuroSDK directory or in the network folder with the corresponding access rights. If you use a standard neural network (training wasn't performed in operating conditions), we guarantee the overall accuracy of 80-95% and the percentage of false positives of 5-20% (see Data collection requirements for neural network training).
Scanning window	Yes	The scanning mode is disabled by default. To enable the scanning mode, select the Yes value
Scanning window	No
Scanning window height	540	The height and width of the scanning window are determined according to the actual size of the frame and the required number of windows. For example, the real frame size is 1920×1080 pixels. To divide the frame into four equal windows, set the width of the scanning window to 960 pixels and the height to 540 pixels
Scanning window step height	540	The scanning step determines the relative offset of the windows. If the step is equal to the height and width of the scanning window, respectively, the segments will line up one after another. Reducing the height or width of the scanning step will increase the number of windows due to their overlapping each other with an offset. This will increase the detection accuracy but will also increase the CPU load Attention! The height and width of the scanning step must not be greater than the height and width of the scanning window. The detector won't operate with such settings.
Scanning window step width	480	The scanning step determines the relative offset of the windows. If the step is equal to the height and width of the scanning window, respectively, the segments will line up one after another. Reducing the height or width of the scanning step will increase the number of windows due to their overlapping each other with an offset. This will increase the detection accuracy but will also increase the CPU load Attention! The height and width of the scanning step must not be greater than the height and width of the scanning window. The detector won't operate with such settings.
Scanning window width	480	The height and width of the scanning window are determined according to the actual size of the frame and the required number of windows. For example, the real frame size is 1920×1080 pixels. To divide the frame into four equal windows, set the width of the scanning window to 960 pixels and the height to 540 pixels
Track lifespan (starting with Detector Pack 3.14)	Yes	By default, the parameter is disabled. If you want to display the track lifespan for an object in seconds, select the Yes value
Track lifespan (starting with Detector Pack 3.14)	No
Basic settings
Mode	CPU	Select the operation mode of the detector and neural filter (see Hardware requirements for neural analytics operation, Selecting Nvidia GPU when configuring detectors). Neural filter operation mode is used only if you select the Yes value in the Neural filter parameter. Attention! It may take several minutes to launch the algorithm on NVIDIA GPU after you apply the settings. You can use caching to speed up future launches (see Optimizing the operation of neural analytics on GPU in Windows OS, Optimizing the operation of neural analytics on GPU in Linux OS). If you specify another processing resource than the CPU, this device carries most of the computing load. However, the CPU is also used to run the detector. For the Person Down Detection and Sitting Person Detection, the accuracy can depend on the selected processor. If the recognition quality deteriorates when you change the processor, we recommend selecting the best values of the detector parameters and configuring perspective empirically (see Sitting person detector, Person down detector). Starting with Detector Pack 3.11, Intel HDDL and Intel NCS aren’t supported. Starting with Detector Pack 3.14, Intel Multi-GPU and Intel GPU 0-3 are supported.
	Nvidia GPU 0
	Nvidia GPU 1
	Nvidia GPU 2
	Nvidia GPU 3
	Intel NCS (not supported)
	Intel HDDL (not supported)
	Intel GPU
	Intel Multi-GPU
	Intel GPU 0
	Intel GPU 1
	Intel GPU 2
	Intel GPU 3
	Huawei NPU
Pose estimation neural network (starting with Detector Pack 3.14)	Pose estimation	Select a pose estimation neural network. The names may indicate the size of the neural network. The size (Nano, Middle, Large) of the neural network indicates the amount of resources consumed. The larger the neural network, the higher the accuracy but the higher the load on the processor
	Pose estimation (Nano)
	Pose estimation (Middle)
	Pose estimation (Large)
Neural network filter
Neural filter	Yes	Neural filter is disabled by default. To use the neural filter to filter out parts of tracks, select the Yes value
Neural filter	No
Neural filter file		Select a neural network file

By default, the entire frame is a detection area. In the preview window, you can specify the custom detection areas and skip areas. To specify areas, right-click the frame and select the required area (see Configuring a skip area, Configuring a detection area).

The following logic is used for this:

if you specify only detection areas, no detection is performed in all other parts of the frame;
if you specify only skip areas, detection is performed in all other parts of the frame.

Note

For convenience of configuration, you can "freeze" the frame. Click the button. To cancel the action, click this button again.
The detection area is displayed by default. To hide it, click the button. To cancel the action, click this button again.
To delete the selected area, click the button.

To save the parameters of the detector, click the Apply button. To cancel the changes, click the Cancel button.

Configuring the Human pose detector is complete. If necessary, you can create and configure the required sub-detectors on the basis of the Human pose detector (see Configuring the Human pose detector sub-detectors).

Page tree

Documentation for Axxon One 2.0. Documentation for other versions of Axxon One is available too.