Closed gd1925 closed 1 year ago
Hello! Thanks for your interest in my code. Several demos are available in the accompanied repositories for my pip packages hsemotion and hsemotion-onnx. For example, the demo for video processing is available here recognize_emotions_video.py. The processing of photos with the PyTorch models is described at test_hsemotion_package.ipynb
Hi, Thank you very much for your reply. The model is super fast indeed. Runs at >30fps on cpu. Thank you for sharing the code base with us!!
Thanks for your positive feedback! Closing the issue
I notice that the linked demos limits the number of faces which can be recognized to one. Did you ever do any implementation which can track multiple faces and keep track of the identity of the different faces over time? E.g. assign a face id for each face which is reliable such that the face_id is not mixed up over time.
I have an example here https://github.com/av-savchenko/face-emotion-recognition/blob/main/src/video_summarizer.ipynb
However, for now, I believe it is better to use a special tracker, e.g. deepsort with face embeddings from insightface. I do not have an example for such an implementation but it is possible to google it
ср, 2 окт. 2024 г., 20:30 timlac @.***>:
I notice that the linked demos limits the number of faces which can be recognized to one. Did you ever do any implementation which can track multiple faces and keep track of the identity of the different faces over time? E.g. assign a face id for each face which is reliable such that the face_id is not mixed up over time.
— Reply to this email directly, view it on GitHub https://github.com/av-savchenko/face-emotion-recognition/issues/43#issuecomment-2389229102, or unsubscribe https://github.com/notifications/unsubscribe-auth/AFZQMVD4LXZP5CY6IB26KVTZZQUURAVCNFSM6AAAAABPIEPGRGVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDGOBZGIZDSMJQGI . You are receiving this because you modified the open/close state.Message ID: @.*** com>
Hi,
Is there a demonstration code available for running videos or webcam feeds? Thank you.