A Palmprint Recognition Method Based on Cross Gradient Coding with Stable Feature Image

Abstract

The invention belongs to the technical field of identification, and relates to a palmprint recognition method of cross-gradient coding under a stable feature image. First, the corner points between the index finger and the middle finger, the ring finger and the little finger in the palmprint image are respectively detected, and the two corners The tangent line formed by the points is rotated and corrected to segment the image of the region of interest of the original palmprint image; then the image of the region of interest is normalized by grayscale to obtain the normalized palmprint image, and an energy functional model is established and solved to obtain Stable feature image, and then perform cross-gradient encoding on the stable feature image to obtain cross-gradient encoding features for palmprint matching and recognition; after the palmprint matching is completed, the matching result is automatically output; the recognition method is simple, the recognition accuracy is high, the algorithm complexity is low, and the recognition time is short. Short, strong anti-interference.

Description

Technical field: [0001] The invention belongs to the technical field of identification, and relates to an identification method based on human biological characteristics, in particular to a palmprint identification method with cross-gradient encoding under a stable feature image. Background technique: [0002] In today's highly informationized society, identification is one of the basic methods to strengthen the security of information and systems. Traditional identification technologies, such as using keys, password locks, ID cards, etc., are inconvenient, unsafe, and dangerous. Reliable and many other shortcomings, and biometric technology is an effective way to overcome these shortcomings. People began to study and design biometric identification technology in 1960. In June 2003, the International Civil Aviation Organization of the United Nations announced its application plan for biotechnology. Biometrics, such as fingerprints, irises, and face recognition, will be added...

AI Technical Summary

Problems solved by technology

Among the above palmprint recognition methods, the method based on the main line extraction is greatly disturbed by external factors, and the foreground and background of the palmprint are not easy to distinguish, so it is difficult to accurately extract the main line; Recognition has better recognition results, but for palmprint images, it lacks the description of texture and other information; the palmprint recognition method based on encoding uses encoding to encode the characteristics of palmprint, which can obtain relatively ideal recognition results, which is relatively The classic and effective methods are the methods listed above. PalmCode and FusionCode are the most cost-effective. They can not only obtain higher recognition accuracy, but also the algorithm complexity is not high compared with subsequent algorithms. However, the above methods have some disadvantages. Defects: In the first instance, filtering is required to smooth the image before feature extraction. The purpose is to reduce noise interference and remove some pseudo-features that affect recognition, but simply using filtering to smooth the image cannot obtain a more stable palm. It is not easy to control the degree of filtering; second, most of them use Gabor transform to extract directional features. Not only is it time-consuming to filter palmprint images, but Gabor filtering is mostly a DC component, and it is difficult to describe palmprint lines. Not the best choice; the third is that the palmprint is easily affected by rotation, translation, etc. during collection, which makes the above method less error-tolerant when using Hamming distance for matching

Images