A novel verification approach for silicon retina stereo matching algorithms

摘要

This paper presents a novel verification approach for silicon retina based stereo matching algorithms. The silicon retina is a new type of bio-inspired sensor technology derived from the human vision system. The sensor only provides event-triggered information and avoids redundant data by delivering intensity changes in a scene. Due to the special characteristics and limitations of this sensor technology, existing verification approaches and performance metrics are not suitable. Thus, a tool for synthetic scene generation, ground truth generation and algorithm verification has been implemented.