Enhancement of Color Images in HOT domain with quantitative measurements using entropy and relative entropy

Color Images captured under complex lighting conditions may suppress the visual cue information. In order to perform consistent analysis from the images they need to be enhanced in terms of brightness along with colors restored with minimal computational overhead. Transform domain image enhancement offers a superior visual brilliance. This paper presents Hirschman Optimal Transform (HOT) based enhancement of color images. The features of HOT are its basis kernel has more no. of zeros and has high frequency resolution. These features are exploited to achieve reduced computational complexity and enhanced image quality in terms of brightness, contrast & color. Further the proposed HOT based enhancement resulted in a better JPEG Quality Metric compared to conventional DCT based transform domain enhancement. The enhanced visual features of the subject images using HOT are found to be superior to the contemporary transform domain methods. We also used the concepts of information theory, entropy relative entropy to evaluate the enhanced images.