3-D color histogram equalization by principal component analysis

• The proposed method uses a one dimensional histogram for color HE.
• The proposed method applies equalization on a transformation of input image.
• The proposed method uses the projection of image on the first component of PCA.
• The resulted images of the proposed method are free from brightening.
• The resulted mean brightness differences are not very significant.

Color histogram equalization is a method for improving visual appearance of images by enhancing image contrast. Color histogram equalization methods are mostly faced with problems like over-enhancement and brightening. In this paper a new color histogram equalization method is proposed which defines a new three dimensional cumulative distribution function based on a one-dimensional histogram. This one-dimensional histogram is calculated by taking into account the correlation between color channels using PCA. Over-enhancement and brightening are solved by this method because of applying the equalization on a transformed image instead of image itself.