New Approach to Edge Detection on Different Level of Wavelet Decomposition

Abstract

This paper proposes a new approach to edge detection on the images over which the wavelet decomposition was done to the third level and consisting of different levels of detail (small, medium and high level of detail). Images from the BSD (Berkeley Segmentation Dataset) database with the corresponding ground truth were used. Daubechies wavelet was used from second to tenth order. Gradient and Laplacian operators were used for edge detection. The proposed approach is applied in systems where information is processed in real time, where fast image processing is required and in systems where high compression ratio is used. That is, it can find practical application in many systems, especially in television systems where the level of details in the image changes. The new approach consists in the fact that when wavelet transform is applied, an edge detection is performed over the level 1 image to create a filter. The filter will record only those pixels that can be potential edges. The image is passed through a median filter that filters only the recorded pixels and 8 neighbors of pixel. After that, the edge detection with one of the operators is applied onto the filtered image. F measure, FoM (Figure of Merit) and PR (Performance Ratio) were used as an objective measure. Based on the obtained results, the application of the proposed approach achieves significant improvements and these improvements are very good depending on the number of details in the image and the compression ratio. These results and improvements can be used to improve the quality of edge detection in many systems where compressed images are processed, that is, where work with images with a high compression ratio is required.