Figure-Ground Processing, Saliency and Guided Attention for Analysis of Large Natural Scenes

Professor Shihab Shamma (ECE/ISR) will contribute to an Office of Naval Research (ONR) MURI, "Figure-Ground Processing, Saliency and Guided Attention for Analysis of Large Natural Scenes." Maryland's portion of the award is $1.3 million. This award was made in MURI category 5: Bio-inspired Autonomous Agile Sensing and Exploitation of Regions of Interest within Wide Complex Scenes. The research will advance the computational architecture of sensors in large acquisition systems so that surveillance tasks in large natural scenes with complex imagery can be better accomplished. Shamma's contribution will be to define regions of interest both spatially and perceptually; to describe how visual and auditory search on the perceptual objects is guided by bottom-up saliency and by top-down knowledge of target features; to to search among these potential targets with a specific goal in mind; and to transfer the knowledge obtained from the neurophysiology, perceptual psychophysics and neural modeling into algorithms and architectures for solving problems of relevance.