Image Segmentation by Edge Detection Technique

Image segmentation is one of the topic of research for last decade. Edge detection in images significantly preserves the important structural properties in an image and reduces the amount of data and filters out useless information. In this paper, the two most commonly used edge detection methods (Laplacian and Sobel edge detectors) are discussed. It is found that Sobel edge detection algorithms perform better than Laplacian algorithms; however, the false edges are high in both cases for blurred or low resolution images. Therefore, a new algorithm and set of filters (kernels) is proposed and its results are compared with the Sobel and Laplacian filters for three images. From the results obtained it is found that the proposed algorithm performs better than in terms of less false edges than the Sobel and Laplacian filters.