Summary
Compatibility of face_recognition_demo with OpenVINO and steps to run it.
Description
Possible errors include:
- Inference of face_recognition_demo with OpenVINO™ 2021 version throws error:
Error:ImportError:DLL load failed while importing ie_api
- Inference of face_recognition_demo with OpenVINO™ 2021 version and Intel® Neural Compute Stick 2 (Intel® NCS2) plugin throws error :
AttributeError: 'openvino.inference_engine.ie_api.IENetwork' object has no attribute 'layers'
Resolution
- Download the following pre-trained models:
- face-detection-retail-0004 and face-detection-adas-0001, to detect faces and predict their bounding boxes;
- landmarks-regression-retail-0009, to predict face key points;
- face-reidentification-retail-0095, to recognize persons
- Execute the face_recognition_demo.
- Setup OpenVINO environment : cd <INSTALL_DIR>\openvino_2020.3.194\deployment_tools\bin\setupvars.bat
- Navigate to the face recognition demo folder : cd <INSTALL_DIR>\openvino_2020.3.194\deployment_tools\open_model_zoo\demos\python_demos\face_recognition_demo
- Run pip install -r requirement.txt
- Run the demo : python./face_recognition_demo.py -m_fd "<INSTALL_DIR>\openvino_2020.3.194\deployment_tools\tools model_downloader\intel\face-detection-retails-0004\FP16\face-detection-retail-0004.xml" -m_lm "<INSTALL_DIR>\openvino_2020.3.194\deployment_tools\tools model_downloader\intel\landmarks-regression-retail-0009\FP16\landmarks-regression-retail-0009.xml" -m_reid "<INSTALL_DIR>\openvino_2020.3.194\deployment_tools\tools model_downloader\intel\face-reidentification-retail-0095\FP16\face-reidentification-retail-0095.xml" -i 0 -fg "<IMAGE_DIR>\Pictures\image_name"
| Note |
Make sure to download the models and execute the demo from the same environment as your OpenVINO™ version. |
Additional information
The face_recognition_demo is only available until OpenVINO™ 2020.3 release. Refer to the Interactive Face Recognition Demo documentation for more details.