search

phoenix104104 / fast_blind_video_consistency

Learning Blind Video Temporal Consistency (ECCV 2018)
http://vllab.ucmerced.edu/wlai24/video_consistency/
Other
423 stars 64 forks source link

Inputs to the model #3

Closed musicrainie closed 6 years ago

musicrainie commented 6 years ago

Thanks for your great works. In this code, the inputs to the model is (frame_p2, frame_o1, frame_i2, frame_o1) and the provided results comply with it, which seems not to be consistent with the paper if I don't misunderstand the details. Is it just the case that frame_i1 is not used as an input to the model, which means it doesn't need to appear at all in the test phase?

phoenix104104 commented 6 years ago

Hi, you are right. The inputs to the model should be (frame_p2, frame_o1, frame_i2, frame_i1). frame_i1 should be used in both training and testing. I will update the code and model.

phoenix104104 commented 6 years ago

Code and model are updated.