Gender recognition using PCA and LDA with improve preprocessing and classification technique

This paper explains the gender recognition system through a human facial image by using the basic method of Principal Component Analysis (PCA) combined with Linear Discriminant Analysis (LDA). PCA+LDA method performance can be improved by improvising the preprocessing techniques such as resizing the image, equalizing the histogram, and removing the variation of the image background by adding oval masking face. Furthermore, in classification process, using 9 nearest neighbors gives the better recognition accuracy rather than using only 1 nearest neighbor. The highest accuracy results obtained with the proposed method is superior to get 89.70% when compared to the PCA + LDA method without adding masking face, which only reached 84.16%.