The Research of Face Recognition Method Based on SparseRepresentation and Feature Selection
暂无分享,去创建一个
Based on the digital image facial recognition technology is as the research background. Based on the analysis of the existing face recognition methods, combining with the latest theory of pattern recognition is for face recognition of facial expression. Illumination such as complicated conditions is in-depth study of face recognition based on sparse representation and feature selection problem. Sparse representation problem of computational complexity is with the increase of the dictionary size increase rapidly. To this end, this paper proposes a fast decomposition gradient projection algorithm in solving sparse representation (FDGP). By minimizing a quadratic programming problem of bounded constraint to solve the problem of sparse representation, the process of gradient projection iteration does not solve the problem, but a select gradient is the biggest change elements as working set, which converts large-scale optimization problems small bounded constrained quadratic programming problem to solve and saves memory consumption. It significantly improves the efficiency of large scale sparse representation problem and ultimately increases the accuracy and efficiency of face recognition.
[1] Frederick I. Parke,et al. Computer generated animation of faces , 1972, ACM Annual Conference.
[2] Allen Y. Yang,et al. Robust Face Recognition via Sparse Representation , 2009, IEEE Transactions on Pattern Analysis and Machine Intelligence.