Classification of lip color based on multiple SVM-RFE

Classification of lip color is an important aspect in the theory of Traditional Chinese Medicine (TCM). The lip color of one person can reflect the person's healthy status. This paper investigates the effectiveness of multiple support vector machine recursive feature elimination (SVM-RFE) for feature selection in the classification of lip color. In the proposed method, both the normalized histogram features and the mean/variance features are computed for the ranking score from a statistical analysis of weight vectors of multiple linear SVMs trained on subsamples of the original training data. Experimental results show that not only the multiple SVM-RFE is effective for feature selection in the lip color classification, but also the accuracy rate of classification of the proposed method is better than the existing SVM method, which is close up to 91%.

[1]  Abdesselam Bouzerdoum,et al.  Skin segmentation using color pixel classification: analysis and comparison , 2005, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[2]  Jason Weston,et al.  Gene Selection for Cancer Classification using Support Vector Machines , 2002, Machine Learning.

[3]  Abdesselam Bouzerdoum,et al.  Adaptive skin segmentation in color images , 2003, 2003 IEEE International Conference on Acoustics, Speech, and Signal Processing, 2003. Proceedings. (ICASSP '03)..

[4]  Sabine Süsstrunk,et al.  Consistent image-based measurement and classification of skin color , 2005, IEEE International Conference on Image Processing 2005.

[5]  Vladimir Vezhnevets,et al.  A Survey on Pixel-Based Skin Color Detection Techniques , 2003 .

[6]  F. Azuaje,et al.  Multiple SVM-RFE for gene selection in cancer classification with expression data , 2005, IEEE Transactions on NanoBioscience.

[7]  Michael Harville,et al.  Automatic Skin Pixel Selection and Skin Color Classification , 2006, 2006 International Conference on Image Processing.

[8]  Wei Li,et al.  Lip color classification based on support vector machine and histogram , 2010, 2010 3rd International Congress on Image and Signal Processing.