UC San Diego

Computer vision is a research field that teaches computers to perform visual tasks such as understanding of a robot’s surroundings or the ability to recognize objects. These are done through the use of sensors and algorithms. In the world today, many algorithms are able to accurately perform complex tasks by emulating various aspects of biological systems. This thesis investigates a novel human-inspired camera system (HICS) which emulates the side-to-side motion of a walking person and evaluates how it affects depth estimation accuracy. It is then demonstrated that the HICS camera improves depth estimation accuracy for monocular and stereo cameras compared to traditional stationary cameras.