Application of color histogram to human face detection
Hassan Fahmy Hashem, Abbas Amin, Ali Al-Sherbiny · 2008
In this paper, we proposed a new face detection method. This new method depends on the color distribution of human faces. The color distribution of human face contains unique characteristics. We used color histograms to form a vector that captures the special relations of these unique regional characteristics by concatenating the color histograms of different regions of human face. Color histogram can be considered as a statistical measure of any image. We also used principle component analysis [PCA] to reduce the dimensionality of the histograms to improve efficiency. Finally we used back propagation neural network to distinguish between face and non-face images.