Skin colour classification using linear discriminant analysis and colour mapping co-occurrence matrix

• Published in: International Conference on Computer Applications Technology, ICCAT 2013
• Main Author: Osman G.; Hitam M.S.
• Format: Conference paper
• Language: English
• Published: 2013
• Online Access: https://www.scopus.com/inward/record.uri?eid=2-s2.0-84879874458&doi=10.1109%2fICCAT.2013.6522047&partnerID=40&md5=d3a5e7e5098459eecb713bf68341fad2

This paper proposed a new technique for region-based skin colour classification using texture information. The texture information was extracted from the colour mapping co-occurrence matrix (CMCM). The thirteen Haralick's textures have been computed and used for formulating a skin colour classifiers using stepwise linear discriminant analysis (LDA). The performance of each skin colour classifier was measured based on true and false positive value. The results shown that the skin colour classifier formulated with [RGB] CMCM at direction (1, 0°) most superior as compared to other direction. Its true positive and false positive are 99.19 percent and 3.83 percent, respectively. Meanwhile, the classifier formulated with [RGB] CMCM at direction (1, 90°) is totally failed to classify skin and nonskin colours. Meaning that, the texture features which are computed from [RGB] CMCM at direction (1, 90°) cannot represent skin and nonskin colour at all © 2013 IEEE.