OpenFace
OpenFace is an opensource face recognition AI model. The algorithm calculates128 measurements, after locating 68 face landmarks, of each face and use this number to identify a known face. The image was created by Brandon Amos who works on OpenFace.
This docker image is built and tested on Firefly-RK3399.
To run the image inside a container. Start the container with:
docker run -it trained-openface bash
Go to the working directory of openface
cd /home/project/openface
Run the classifier with the expected arguments i.e. .pkl file and the test image
./demos/classifier.py infer ./generated-embedings/classifier.pkl test_images/kurt/kurt_4.jpg
To run OpenFace without going inside the container:
docker run -it trained-openface bash -l -c '/home/project/openface/demos/classifier.py infer /home/project/openface/generated-embeddings/classifier.pkl /home/project/openface/test_images/kurt/kurt_4.jpg
Comments
No customer comments for the moment.