G-API Gesture Recognition Demo¶
This demo demonstrates how to run Gesture (e.g. American Sign Language (ASL) gestures) Recognition models using OpenVINO toolkit.
How It Works¶
The demo application expects a gesture recognition model in the Intermediate Representation (IR) format.
As input, the demo application takes:
a path to a video file or a device node of a webcam specified with a command line argument
a path to a file in JSON format with gesture class names
The demo workflow is the following:
The demo application reads video frames one by one, runs person detector that extracts ROI, tracks the ROI of very first person. Additional process is used to prepare the batch of frames with constant framerate.
Batch of frames and extracted ROI are passed to artificial neural network that predicts the gesture.
The app visualizes results of its work as graphical window where following objects are shown:
Input frame with detected ROI.
Last recognized gesture.
By default, Open Model Zoo demos expect input with BGR channels order. If you trained your model to work with RGB order, you need to manually rearrange the default channels order in the demo application or reconvert your model using the Model Optimizer tool with the
--reverse_input_channels argument specified. For more information about the argument, refer to When to Reverse Input Channels section of Embedding Preprocessing Computation.
Creating a Gallery for Gestures Window¶
The gallery of sample videos and list of paths to gesture videos must be created to show the sample gestures on an additional window:
Put videos containing gestures to a separate empty folder. Each video must have only one gesture.
python3 <omz_dir>/demos/gesture_recognition_demo/cpp_gapi/create_list.py --classes_map <path_to_a_file_with_gesture_classes> --gesture_storage <path_to_directory_with_gesture_videos>command, which will create a
gesture_gallery.jsonfile with list of gestures and paths to appropriate videos.
Preparing to Run¶
For demo input image or video files, refer to the section Media Files Available for Demos in the Open Model Zoo Demos Overview. The list of models supported by the demo is in
<omz_dir>/demos/gesture_recognition_demo/cpp_gapi/models.lst file. This file can be used as a parameter for Model Downloader and Converter to download and, if necessary, convert models to OpenVINO IR format (*.xml + *.bin).
An example of using the Model Downloader:
omz_downloader --list models.lst
An example of using the Model Converter:
omz_converter --list models.lst
Refer to the tables Intel’s Pre-Trained Models Device Support and Public Pre-Trained Models Device Support for the details on models inference support at different devices.
Running the application with the
-h option yields the following usage message:
gesture_recognition_demo_gapi [OPTION] Options: -h Show this help message and exit. -m_a Required. Path to an .xml file with a trained gesture recognition model. -m_d Required. Path to an .xml file with a trained person detector model. -i Required. Path to a video file or a device node of a webcam. -o Optional. Name of the output file(s) to save. -limit Optional. Number of frames to store in output. If -1 is set, all frames are stored. -c Required. Path to a file with gesture classes. -s Optional. Path to a directory with video samples of gestures. -t Optional. Threshold for the predicted score of an action. -d_d "<device>" Optional. Target device for Person Detection network (the list of available devices is shown below). -d_a "<device>" Optional. Target device for Gesture Recognition (the list of available devices is shown below). -no_show Optional. Don't show output. -u Optional. List of monitors to show initially.
Running the application with an empty list of options yields an error message.
For example, to do inference on a CPU, run the following command:
./gesture_recognition_demo_gapi -m_a <path_to_model>/asl-recognition-0004.xml \ -m_d <path_to_model>/person-detection-asl-0001.xml \ -i 0 \ -c <omz_dir>/data/dataset_classes/msasl100.json
Run-Time Control Keys¶
The demo starts in person tracking mode and to switch it in the action recognition mode you should press
0-9 button with appropriate detection ID (the number in top-left of each bounding box). If frame contains only one person, they will be chosen automatically. After that you can switch back to tracking mode by pressing space button.
An example of file with class names can be found within the OMZ directory:
NOTE : To run the demo application with video examples of gestures specify the
-skey with valid path to the directory with video samples. The name of each video sample should be the valid name of gesture from
<omz_dir>/data/dataset_classes/msasl100.jsonfile. To navigate between samples use ‘f’ and ‘b’ keys for iterating next and previous video sample, respectively.
You can save processed results to a Motion JPEG AVI file or separate JPEG or PNG files using the
To save processed results in an AVI file, specify the name of the output file with
aviextension, for example:
To save processed results as images, specify the template name of the output image file with
pngextension, for example:
-o output_%03d.jpg. The actual file names are constructed from the template at runtime by replacing regular expression
%03dwith the frame number, resulting in the following:
output_001.jpg, and so on. To avoid disk space overrun in case of continuous input stream, like camera, you can limit the amount of data stored in the output file(s) with the
limitoption. The default value is 1000. To change it, you can apply the
-limit Noption, where
Nis the number of frames to store.
> NOTE : Windows* systems may not have the Motion JPEG codec installed by default. If this is the case, OpenCV FFMPEG backend can be downloaded by the PowerShell script provided with the OpenVINO install package and located at
<INSTALL_DIR>/opencv/ffmpeg-download.ps1. Run the script with Administrative privileges. Alternatively, you can save results as images.
The application uses OpenCV to display gesture recognition result and current inference performance.