A Novel Face Recognition Model Based on Feature Registration

With the rapid development of information construction, identity identification has become an important factor affecting the economic and social security in the era of big data. Face recognition is the core research direction in the field of identity authentication and image processing security. In recent years, the research of face recognition mainly focuses on how to eliminate the influence of environmental factors and facial occlusion. A face detection model based on feature registration is proposed, which uses Gaussian filter processing to realize image noise reduction, and introduces the improved Prewitt operator edge detection to realize binarization. Through the gray level threshold separation and screening of specific areas, on the basis of position determination, location of face organs and parameters were analyzed, and the face recognition model was constructed, and puts forward the improved Harris corner feature extraction methods in combination with Normalized Cross Correlation face matching normalization algorithm. The experimental effect showed that the experimental accuracy reached 90.13%, and time efficiency was increased by 6.67%.