Quang Tran
Results
1
comments of
Quang Tran
According to the paper, I think it does not follow the adversarial design. Instead, it uses a single model for matting, incorporating three major modules (encoder, decoder, and deep guided...