改进的PCA-LDA人脸识别算法的研究

Research on Improved PCA-LDA Face Recognition Algorithm

主成分分析算法(PCA)和线性鉴别分析算法(LDA)被广泛用于人脸识别技术中,但是PCA由于其计算复杂度高,致使人脸识别的实时性达不到要求。线性鉴别分析算法存在“小样本”和“边缘类”问题,降低了人脸识别的准确性。针对上述问题,提出使用二维主成分分析法(2DPCA)与改进的线性鉴别分析法相融合的方法。二维主成分分析法提取的特征比一维主成分分析法更丰富,并且降低了计算复杂度。改进的线性鉴别分析算法重新定义了样本类间离散度矩阵和Fisher准则,克服了传统线性鉴别分析算法存在的问题,保留了最有辨别力的信息,提高了算法的识别率。实验结果表明,该算法比主成分分析算法和线性鉴别分析算法具有更高的识别率,可以较好地用于人脸识别任务。

Principal component analysis(PCA)and linear discriminant analysis(LDA)are widely used in face recognition technology.However,the computational complexity of PCA is so high that the real-time performance of face recognition cannot meet the requirements.LDA has“small samples”and“edge”problems,which reduces the accuracy of face recognition.In view of the above problems,we propose a method that fuses two-dimensional principal component analysis and improved linear discriminant analysis.The features extracted by 2DPCA are better,faster than PCA,and the computation time is reduced.The improved LDA redefines the dispersion matrix and Fisher criterion between samples,overcomes the problems of the traditional LDA algorithm,retains the most discerning information,and enhances the recognition rate of the algorithm.The experiment shows that the proposed algorithm has a higher recognition rate than the principal component analysis and linear discriminant analysis,which can be better used for face recognition tasks.