A Novel algorithm of local contrast enhancement for medical image

Abstract

The local contrast enhancement of medical image is useful and important to the medicine diagnosis. The current techniques of local enhancing exists both over-enhancing and high complexity problems. The over-enhancement may cause the clinical information loss. In this paper, we propose the weighted calculation to prevent the unexpected effects appearing and the local bihistogram equalization (LBHE) to reduce the over-enhancing artifacts. The constraint curve provides a mechanism of automatic modification. We also propose the pixel-matching scheme for similar value search and an interlaced data flow for implementing our algorithm in parallel. The experimental results have proven that our method enhances efficiently the adjacent organic structure of the heart in MRI without local noise.