First, thank you to release your code in public to used , I want ask you about form in txt file , which means if you can upload a copy for this file to applied the same on own dataset, please ?
Moreover, Can we using landmark that got it from https://github.com/mpc001/Visual_Speech_Recognition_for_Multiple_Languages
when crop_mouth to continue our work in https://github.com/mpc001/Lipreading_using_Temporal_Convolutional_Networks ?
or should landmark be from dilb method? because I faced problem in dimantion
regards