Somar Boubou

Toyota Technological Institute, Kyushu University

Papers

Total Citations

H-Index

About

Somar Boubou is a robotics researcher focused on real-time visual perception and autonomous navigation. His work centers on enabling robots to perceive, track, and interact with their environment using efficient computer vision algorithms. A key contribution is the development of a real-time system for person recognition and pursuit using 3D depth data, achieving 8 citations for its practical approach to human-robot interaction. Earlier, Boubou implemented the Camshift algorithm on a mobile robot for color-based person tracking with a monocular camera, demonstrating computationally efficient tracking that remains relevant for real-time applications (6 citations). He also advanced visual localization for autonomous robots by introducing "visual impressions"—HSV color distributions stored in clustering feature trees—to enable place recognition without complex maps (4 citations). This novel approach weights color entries to improve localization accuracy. Boubou’s work bridges efficient algorithms and practical robotics, offering scalable solutions for tracking and navigation that are particularly valuable for students and researchers developing low-cost, vision-based autonomous systems.