Hi all.
I need to evaluate my tracking system how accurate it works. So I need a face video dataset that contains multiple faces in each frame and also the labels contain bounding boxes to represent the coordinate of faces. The format of labels should be like this:
frame1, track_id1, face1 , [x_left_1, y_left_1, w1, h1], face2, track_id2, [x_left_2, y_left_2, w2, h2]
Briefly, I need a face video dataset for face tracking with label files that contain track_id and boxes of faces appearing in each video frame. I mean the dataset should support multiple faces in each frame.
Dear Saeedeh Bahrami
These articles might be useful, have a look:
1. https://www.mygreatlearning.com/blog/real-time-face-detection/
2. https://lionbridge.ai/datasets/20-best-bounding-box-image-and-video-datasets-for-machine-learning/
3. https://www.cs.tau.ac.il/~wolf/ytfaces/
Kind Regards
Qamar Ul Islam
Thanks a lot, Qamar Ul Islam for sending some links.
I read the above links. But the first one is for face detection. The YouTube Faces dataset only supports a single face in each frame.
I need a face video dataset for face tracking with label files that contain track_id and boxes of faces appearing in each video frame. I mean the dataset should support multiple faces in each frame.
here is a tutorial for python that can help you
https://www.youtube.com/watch?v=R7B8sSByZGQ
- Badges
- Science topic