Medical image enhancement strategy based on morphologically processing of residuals using a special kernel

Abstract

Medical imaging is playing a pivotal role in the domain of medical diagnosis. But the major problems related to medical images is that most of the images (of different modalities) are suffered from noise and other quality-related problems such as poor contrast, blurring, and difficulties in extracting appropriate information. Therefore, it is necessary to construct some techniques that could improve medical images in such a way that it will be ideal for diagnostic applications. Medical image enhancement is the process of increasing the contrast quality of intensity variations and improves the visual representation of medical images. To provide suitable interpretation and clearer image for the observers with reduced noise levels, this article proposes a novel competent medical image enhancement method based on morphologically processing of residuals using a special kernel. First, it combines linear low pass filtering with nonlinear technique that allows for selection of essential regions where edges will get well preserved. The selection of those regions is based on morphological processing of linear filter residuals and then aims to find significant regions specified by edges of high amplitude and appropriate size. The reconstructed regions are combined with the output of low pass filtering to recovers the original shape of edges. In addition, the method allows to control the contrast to avoid blurring while preserving significant image information. In the end, a special kernel is convolved with an image to obtain the sharper image. Experiments on the different types of medical images are performed and the results are compared using different standard evaluation metrics. The quantitative results obtained through proposed method for all the metrics are optimal compared to other competing methods. It is observed that the results on various test images affirm that the proposed method generates excellent image visual quality and significantly enhances the contrast of all images which helps with better diagnosis and treatment.