Learning what and where to attend
This work addresses the need for better attention mechanisms in computer vision by providing stronger supervisory signals, though it is incremental as it builds on existing attention networks.
The paper tackled the problem of improving visual recognition by teaching deep convolutional networks to attend to image regions humans find important, using a new dataset of human attention maps. The result was a significant improvement in network accuracy and more interpretable features.
Most recent gains in visual recognition have originated from the inclusion of attention mechanisms in deep convolutional networks (DCNs). Because these networks are optimized for object recognition, they learn where to attend using only a weak form of supervision derived from image class labels. Here, we demonstrate the benefit of using stronger supervisory signals by teaching DCNs to attend to image regions that humans deem important for object recognition. We first describe a large-scale online experiment (ClickMe) used to supplement ImageNet with nearly half a million human-derived "top-down" attention maps. Using human psychophysics, we confirm that the identified top-down features from ClickMe are more diagnostic than "bottom-up" saliency features for rapid image categorization. As a proof of concept, we extend a state-of-the-art attention network and demonstrate that adding ClickMe supervision significantly improves its accuracy and yields visual features that are more interpretable and more similar to those used by human observers.