Gal Sela

Papers

Total Citations

H-Index

About

Gal Sela is a researcher specializing in robotic vision and active perception systems, with a particular focus on attention mechanisms and biologically-inspired visual processing. His most influential work, the 1997 thesis "Real-Time Attention for Robotic Vision," garnered 51 citations and laid important groundwork in the field by developing and implementing algorithms for detecting interest points in visual scenes. Drawing inspiration from primate visual systems, Sela's approach employed nonuniform sampling resolution to enable robot-mounted cameras to dynamically focus attention — a significant contribution to the challenge of real-time robotic perception. Building on this foundation, his 2002 paper on fast computation of multiscalar symmetry in foveated images further advanced the development of active vision systems for mobile robots. This work introduced an overlapping receptive field model for efficient data reduction, combining a biologically-inspired foveated vision sensor with adapted scan-line algorithms to achieve practical computational performance. Sela's research sits at the intersection of computer vision, robotics, and computational neuroscience. His contributions helped bridge biological models of visual attention with real-world robotic applications, influencing subsequent work in autonomous systems and active vision architectures.