Improving the Performance of Machine Learning Based Multi Attribute Face Recognition Algorithm Using Wavelet Based Image Decomposition Technique | Science Publications

Journal of Computer Science

M. Rajaram¹  S. Sakthivel¹ 

关键词: Multi resolution analysis;    support vector machine;    wavelet transformation;    Discrete Cosine Transformation (DCT);    Principal Component Analysis (PCA);    Linear Discriminant Analysis (LDA);    Multi-Dimensional Scaling (MDS);    Discrete Wavelet Transform (DWT);    Linear Discriminant Analysis (LDA);    Radial Basis Function (RBF);   

DOI  :  10.3844/jcssp.2011.366.373

学科分类：计算机科学（综合）

来源: Science Publications

【 摘 要 】

Problem statement: Recognizing a face based attributes is an easy task for a human toperform; it is closely automated and requires little mental effort. A computer, on the other hand, has noinnate ability to recognize a face or a facial feature and must be programmed with an algorithm to doso. Generally, to recognize a face, different kinds of the facial features were used separately or in acombined manner. In the previous work, we have developed a machine learning based multi attributeface recognition algorithm and evaluated it different set of weights to each input attribute andperformance wise it is low compared to proposed wavelet decomposition technique. Approach: In thisstudy, wavelet decomposition technique has been applied as a preprocessing technique to enhance theinput face images in order to reduce the loss of classification performance due to changes in facialappearance. The Experiment was specifically designed to investigate the gain in robustness againstillumination and facial expression changes. Results: In this study, a wavelet based imagedecomposition technique has been proposed to enhance the performance by 8.54 percent of thepreviously designed system. Conclusion: The proposed model has been tested on face imageswith difference in expression and illumination condition with a dataset obtained from face imagedatabases from Olivetti Research Laboratory.

【 授权许可】

Unknown