The Optimization-Design of 2DPCA Image Feature Extraction Algorithm Based on Convex Optimization

Abstract

The 2DPCA algorithm has the problems of many retention coefficients and low efficiency. This paper proposes to optimize the 2DPCA objective function based on the convex optimization outer point algorithm, and seeks the optimal scattering degree standard vector series. The experiment results shows, both the 2DPCA method before and after improvement can achieve higher image feature extraction efficiency, but the optimized algorithm has higher recognition accuracy when time is constant, and the recognition accuracy of the optimized algorithm will increase with time consumption.