Edge Detection in Medical Images – Smoothening Techniques

This paper presents the mathematical techniques to improve the performance of edge detection in medical images. Here, we have tried the Topological Derivative, which takes care of the domain features in the medical image for enhancing the edge details by perturbing the domain of interest using a hole or incision. The movement of perturbation is calculated through a cost function, which identifies the affected regions in a medical image up to the cell level of the diseased organ. The accuracy of edge detection is further improved by the Level Set method, which helps in controlling the boundaries strategically by handling the changes in topology for the precise cornering of the edges. Further, Contour let Transformation is used to represent both the contour and the edges in a smoother way. Finally, sparse representation of the data going hand in hand with the Contourlet transformation, along with that of the interpolation methods, which helps to resolve the edges in the image. As a consequence of the comprehensive treatment of these three techniques, the resolution of edge detection in medical images will be improved.