Muti-scale image enhancement based on properties of human visual system

Abstract

The logarithmic image processing (LIP) model is a mathematical framework which has been proved to be consistent with several laws and fit characteristics of the human visual system. In this paper, we both utilize this LIP model and consider characteristics of the human visual system (HVS) to propose a new multi-scale enhancement algorithm. Then a new measure of enhancement based on JND model (Just Noticeable Difference, JND) of human visual system is proposed and used as a tool for evaluating the performance of the enhancement technique. Finally, the proposed algorithm's performance is compared quantitatively to several popular image enhancement algorithms, and experimental results show that the propose algorithm can adjust the image dynamic range, enhance the image details and achieve a more pleasing and comfortable image.