Study on Image recognition based on computer visual angle point detection

Abstract

Considering the objective change of human beings and the effect of temporal parallax, it is difficult to extract the required feature points accurately and recognize the recognition algorithm. In order to obtain effective special feature points, the depth perception mechanism of HVS is simulated, and the robustness of the recognition system is improved by corner detection of facial images. Three-dimensional imaging effect is simulated by two-dimensional image processing method, and image matching is carried out by sparse learning, clustering and other algorithms to reduce error samples and complete identification test.