Edge Detection Scheme Based on the Multi-Resolution form of Singular Value Decomposition

This work presents an edge detection scheme based on the Multi-Resolution form of Singular Value Decomposition (MR-SVD) and compared with Sobel, Prewitt and Roberts cross edge detectors, using synthetic and natural images. The behaviour of the algorithms in noisy environment is demonstrated. Potential fields of application and advantages of presented method are discussed, ideas for following research are highlighted.