High dynamic range for contrast enhancement

Abstract

Video image quality improving technologies produce remarkable achievements as display devices make rapid progress. However, there exist some limitations on intensity representation of display and acquisition devices to reproduce the real world video images. Various contrast enhancement (CE) algorithms have been developed and applied to overcome those limitations but there are many drawbacks such as color information change, loss of the lighting information, and excessive enhancement. This paper proposes a CE algorithm based on high dynamic range (HDR) imaging. First, an input image is determined whether or not it requires CE in a pre-processing step using the proposed auto exposure algorithm. Then, multiple images are generated by applying the intensity mapping function to an input image. At last, an HDR image is constructed with multiple images, in which registration of multiple images is not required. So, the proposed CE algorithm could increase the dynamic range and thus increase the contrast of an input image. The proposed CE method makes use of different intensity information from multiple images, therefore, the performance of the proposed method is better than that of existing CE algorithms. Experiments with three real images show the effectiveness of the proposed CE algorithm