NeurIPS 2019
Sun Dec 8th through Sat the 14th, 2019 at Vancouver Convention Center
Paper ID:91
Title:DeepUSPS: Deep Robust Unsupervised Saliency Prediction via Self-supervision


		
The paper presents a new approach DeepUSPS based on self-supervision and prediction consistency for saliency estimation. Results are very good and similar to supervised methods. At the beginning reviewers were no completely convinced on the goodness of the model which integrates several handcraft methods, but they agree on the fact that rebuttal was satisfying. Thus reviewers found consensus in the acceptance and the area chair agrees too-