Comparative Analysis of Partial Differential Equation-based and Graph Based Methods for Image Segmentation

摘要

In computer vision systems, image segmentation is an important one as far as preprocessing step is concerned. In literature we have several segmentation algorithms, but with constraints due to high processing load and many manualtuning parameters. The challenge in image segmentation in images is inhomogenity levels in intensity. Existing segmentation algorithms depends on region based and it features similarity of intensities of images in the region of interest. The segmentation results are not accurate due to the presence of non uniformity in intensities of images. This paper covers segmentation techniques to overcome the intensity inhomogenity with the help of partial differential equation based methods. This paper critically reviews some of these techniques. It also addresses the quantitative evaluation of segmentation results.