A Study of Pathological Image Detail Enhancement Method Based on Improved Single Scale Retinex

摘要

The pathological image plays an important role in the modern medical and the computer-aided diagnosis. The image enhancement is one critical step in the field of medical image processing and is necessary before image segmentation and image classification. In this paper, a new pathological image enhancement method based on the improved single-scale Retinex (SSR) algorithm is proposed. Firstly, one simple preprocessing including Gaussian filtering (GF) and Gamma correction is performed on the pathological image stained by Haematoxylin and Eosin (H&E). Then the image is enhanced by the improved SSR algorithm in which the GF is replaced by the guided image filtering (GIF) to estimate illumination and the color correction procedure is added to solve the color distortion problem occurred in the previous step. Finally, the enhanced image is acquired after completing a new image detail enhancement process. The experiments with 80 pathological images and the comparison with related work demonstrate that the performance of our proposed method is more satisfactory and more significant.