Improvement of Image Contrast with Local Adaptation

Most existing global contrast enhancement algorithm can not deal with high dynamic range images which are often under non-uniform illumination. Thus a simplified logarithmic transformation equation has been proposed in this paper to meet this demand. Furthermore a partially overlapped sub-block method is also adopted to perform local contrast enhancement. For purpose of optimization, two contrast measures are used for finding optimal transformation parameters. They merge into a new algorithm, called POSLT (Partially Overlapped Sub-block Logarithmic transformation). POSLT exhibits good performance and Controllability even if it has high computational complexity. Our experimental results show its local adaptation for contrast enhancement, the results give firmly support to POSLT.