Problem with extract-mouth-batch.py

rizkiarm / LipNet

Keras implementation of 'LipNet: End-to-End Sentence-level Lipreading'

MIT License

627 stars 224 forks source link

Problem with extract-mouth-batch.py #113

Closed jainnimish closed 2 years ago

jainnimish commented 3 years ago

When I try to train my own model and extract-mouth from the speaker's video, it doesn't correctly predict the face. For example, when I try to predict all the videos inside the evaluations/samples/GRID, it takes the images of the upper right corner of the face instead of the mouth. Please tell me how to fix it and why this is happening.

Any help would be greatly appreciated. Cheers!

Mr-Usama commented 2 years ago

In videos.py use these lines of code will work, using cv2 resize method occurs this problem . from skimage.transform import resize resized_img = resize(frame, new_img_shape)

Mr-Usama commented 2 years ago

When I try to train my own model and extract-mouth from the speaker's video, it doesn't correctly predict the face. For example, when I try to predict all the videos inside the evaluations/samples/GRID, it takes the images of the upper right corner of the face instead of the mouth. Please tell me how to fix it and why this is happening.

Any help would be greatly appreciated. Cheers!

does your project works ?