Bio-inspired visual motion saliency estimation for small video objects with applications

Associate Professor Gang Wang
Beijing Institute of Basic Medical Sciences, China

Detecting small objects against complex backgrounds is highly required yet extremely challenging. Traditional methods mainly characterize the appearance features or the spatially contextual information, suffering from high false alarms and missing rates. The biological visual system has an astonishing ability to quickly select visually important regions in its visual field, using visual motion information to lead the object detection process. Inspired by the mechanism of the biological retina and primary visual cortex, we propose to establish a computational model to simulate the magnocellular pathway, computing the preliminary visual motion saliency cheaply. The spatial-temporal saliency information can help enhance tiny-size objects and faintly discernible targets. We further apply the visual motion computational model as a plug-and-play module to advanced object detection methods. Experimental results have validated the superiority of the proposed scheme.

Biosketch

Gang Wang is an associate professor with the Center of Brain Sciences, Beijing Institute of Basic Medical Sciences, China, a Young Professor with the Chinese Institute for Brain Research, Beijing, China, and an Associate Professor with the University of Electronic Science and Technology of China. He is also serving as the Deputy Secretary-General of Chinese Neuroscience Society, Brain-inspired Intelligence Branch. He received the Ph.D. degree from Ghent University, Belgium in 2019. He is the first/corresponding author of more than 30 papers published in conference proceedings and journals (TPAMI, TIP, CVPR, ICCV, NeurIPS, etc.). He won the CVPR'23 Anti-UAV Challenge, and obtained the best student paper nomination prizes at the EUSFLAT'17 and BNAIC'19. He has received the “Young Talent on Science and Technology” fund from the Chinese government, the “Beijing Nova Program” fund from the Beijing government and the “Young Scholar” fund from CIBR, Beijing. His research interests include computer vision and brain-inspired vision computing.