in this repo we use dlib and resnet model for face verification.
- first we extract the face landmarks and from that we extract the ROI
- then we use the resnet model for extracting face embeddings, which is a 128 length array
- to verify two faces, the embeddings is compared using cosine similarity. if the similarity is greater than 0.9 the faces are match else its not
the code for testing the algorithm is in
main.ipynb
got the models from here : https://github.com/ageitgey/face_recognition_models/tree/master/face_recognition_models/models
Your suggestions and queries matters :)