A novel mathematical morphology based edge detection method for medical images

Abstract

Today, digital image processing is an important area of research. Many medical applications apply image processing techniques to enhance and improve the quality of medical diagnosis. Effective edge detection of images is a fundamental problem in digital image processing. There are different approaches proposed to detect edges of the image such as classical gradients and second derivatives based operators. However, existing approaches have various tradeoffs in the presence of noise and suffer from poor de-noising which leads to an inaccurate edge detection. In this paper, an edge detection approach based on mathematical morphology is proposed to improve edge detection in noisy medical images of fractured bones, head CT, mammogram and colored eye’s fundus. Our evaluation results showed that the proposed detector outperforms classical detectors as well as existing mathematical morphology approaches in detecting edges of medical images under noise effect named “salt and pepper”. The proposed detector applies multiple structured elements to suppress noise and detect edges, thus produces thinner and better linked edges in noisy medical images.