According to different working environment, robots have to configure different sensors. One of the most important is vision sensor. The visual perception is capturing and processing images or image sequences of around environment by vision sensors and computers. Currently, sparse representation and optimization is hot research point in the visual perception area. The presentation of Compressed Sensing (CS) theory makes the problem solving by sparse representation and optimization very popular. Some issues are introduced orderly, such as, image reconstruction, image fusion, face/object recognition, feature extraction and dimension reduction, etc. Besides, the paper summarizes some important algorithms proposed by scientists and engineers at home and abroad. Finally, the mathematical frameworks for above issues are established. The paper also proposes some new methods about image fusion based on online dictionary learning, object recognition using discriminative and collaborative representation, sparse preserving projections for dimensional reduction.
Unfortunately, ACM prohibits us from displaying non-influential references for this paper.
To see the full reference list, please visit http://dl.acm.org/citation.cfm?id=3094251.