Learned Region Sparsity and Diversity Also Predicts Visual Attention
Zijun Wei · Hossein Adeli · Minh Hoai Nguyen · Greg Zelinsky · Dimitris Samaras

Tue Dec 6th 06:00 -- 09:30 PM @ Area 5+6+7+8 #159 #None

Learned region sparsity has achieved state-of-the-art performance in classification tasks by exploiting and integrating a sparse set of local information into global decisions. The underlying mechanism resembles how people sample information from an image with their eye movements when making similar decisions. In this paper we incorporate the biologically plausible mechanism of Inhibition of Return into the learned region sparsity model, thereby imposing diversity on the selected regions. We investigate how these mechanisms of sparsity and diversity relate to visual attention by testing our model on three different types of visual search tasks. We report state-of-the-art results in predicting the locations of human gaze fixations, even though our model is trained only on image-level labels without object location annotations. Notably, the classification performance of the extended model remains the same as the original. This work suggests a new computational perspective on visual attention mechanisms and shows how the inclusion of attention-based mechanisms can improve computer vision techniques.

Author Information

Zijun Wei (Stony Brook)

I am currently a graduate student at Department of Computer Science in Stony Brook Univeristy. From fall 2014 I'm working in the Computer Vision Lab under the supervision of Prof. Dimitris Samaras, Prof. Minh Hoai and Prof. Gregory Zelinsky Prior to this, I received my master degree from the Robotics Institute, Carnege Mellon University in 2013 advised by Prof. Mel Siegel. I work on research problems in Computer Vision and Machine Learning. I am especially interested in plugging human visual perception experience into computer vision to either boost performance or enable human-like results. I am also interested in the other way around -- using computer vision algorithms to model human visual perception systems. I'm a recipient of the Renaissance Technologies Fellowship from 2014 to 2017. I worked as research intern at Adobe Research twice: 2017 spring and 2018 summer.

Hossein Adeli (Stony Brook University)
Minh Hoai Nguyen (Stony Brook University)
Greg Zelinsky (Stony Brook University)
Dimitris Samaras (Stony Brook University)

More from the Same Authors