Garbis Salgian

Papers

Total Citations

H-Index

About

Garbis Salgian is a computer vision researcher whose work focuses on enabling robots and autonomous systems to perceive and navigate complex environments. His key research areas include pedestrian detection, visual odometry, and terrain reconstruction for both ground and underwater robots. Salgian’s most notable contribution is a long-range pedestrian detection system that combines stereo vision with a cascade of convolutional network classifiers, achieving robust detection while reducing false positives and computational load—a foundational approach cited 8 times. He also developed a multi-resolution correlation algorithm for estimating egomotion and terrain structure from image sequences, enabling outdoor robots to map challenging terrain without GPS. His work on complex terrain mapping integrates multi-camera visual odometry with real-time drift correction, allowing low-cost sensor packages to explore GPS-denied regions. These contributions, though modest in citation count, demonstrate Salgian’s practical, systems-oriented approach to solving real-world perception problems in robotics, with applications ranging from autonomous navigation to search-and-rescue operations.