Living body human face detection method based on gray scale symbiosis matrixes and wavelet analysis

Abstract

The invention discloses a living body human face detection method based on gray scale symbiosis matrixes and wavelet analysis. The method comprises: first of all, converting an RGB image comprising a human face area, which is obtained from a camera, into a gray scale image, compressing a gray scale grade to 16 grades, then respectively calculating four gray scale symbiosis matrixes (taking a distance of 1, and angles of 0 degree, 45 degrees, 90 degrees and 13 degrees respectively ), then extracting four texture characteristic quantities including energy, entropy, moment of inertia and correlation on the basis of the gray scale symbiosis matrixes, and respectively obtaining a mean value and a variance for the four texture characteristic quantities of the four gray scale symbiosis matrixes; at the same time, performing secondary decomposition on an original image by use of a Haar small wavelet base, extracting the coefficient matrixes of sub-bands HH1 and HH2 and obtaining a mean value and a variance; and finally sending all characteristic values as samples to be detected to a trained support vector machine for detection, and performing classification identification on real or counterfeit face images. The method provided by the invention has the advantages of reduced calculating complexity and improved detection accuracy.

Description

technical field [0001] The invention relates to face recognition technology in the field of biological pattern recognition, in particular to a living face detection method based on gray-scale co-occurrence matrix and wavelet analysis in a face identity authentication system. Background technique [0002] In recent years, biometric recognition technology has achieved rapid development, and fingerprint recognition, face recognition, iris recognition, etc. have been widely used in identity authentication and other aspects. Among them, face features are widely welcomed because they are in line with the way humans distinguish different people. Compared with traditional keys, passwords, passwords, etc., they are not easy to forge, lose, steal, and are user-friendly. . At present, face recognition technology has been widely used in access control systems, access security checks, case detection, banking systems and other fields. As an effective identity authentication method, face...

AI Technical Summary

Problems solved by technology

The 3D model can imitate various actions such as blinking, speaking, and head movement by modeling the 3D face of a real person. Compared with photos and videos, it poses a greater threat to deception, but forging a living 3D model also has certain risks. difficulty

[0008] 1. Based on motion information analysis: such as judging whether it is a real face based on the 3D depth information of the head; such as using the linear optical flow method to capture the subtle motion information of the 3D face, analyzing the movement of the lips when making a sound, and the eye area The change of head movement, etc., this method has the disadvantages that the features required for liveness detection are difficult to obtain, the user's active cooperation is required, and the computer is expensive;

[0009] 2. Based on the analysis of living body feature information: such as analyzing the thermal infrared image of the face, blinking eyes and other living body features, these methods require additional equipment, have inherent instability factors, and cannot be widely promoted on ordinary equipment;

[0010] 3. Based on texture information analysis: analyze the difference between real faces and photo faces, and use the method of Fourier spectrum and texture features, which may require a large calculation time and space overhead

Images