基于边缘检测与最小二乘支持向量机的人脸图像识别

Research on the Human Faces Image Recognition Based on Edge Detection and Least Square Support Vector Machine

摘要:

文章建立了完整的人脸识别系统,通过图像预处理、特征提取和识别3个过程,实现了人脸的分类识别；研究了图像预处理过程中的边缘检测方法,比较了基于一阶导数的Robert算子、Prewitt算子、Sobel算子和基于二阶导数的Gauss-Laplacian算子的检测效果?在识别阶段采用最小二乘支持向量机方法,进行人脸识别实验,并对实验结果进行比较。

Abstract:

An integrative system of human faces recognition was built up, and human faces recognition experiments were carried out through preprocessing, character extraction, classification and recognition. Methods of edge detection during preprocessing were studied, and each effect of edge detection of Robert operator, Prewitt operator, Sobel operator based on first-order derivative and Gauss-Laplacian operator based on second-order derivative were studied. Least square support vector machine was used during recognition. The results were intercompared.