Learning a time-dependent master saliency map from eye-tracking data in videos | Read Paper on Bytez