Wavelet Difference Reduction Coding with Edge Detection and Tracking

Xu Yan, Fuhai Li, Dong Cheng, Guilin Wen

Abstract: In this paper, we propose a wavelet lossy image coding algorithm with edge detection and tracking using wavelet difference reduction. Firstly, we obtain the edge map of an obstacle image by SUSAN edge detector. And an edge tracking filtering method is applied to the edge information image and Region of Interest (ROI) is extracted from the original image. Then the coordinates of ROI are mapped to the correspondent position in the wavelet domain by ROI mask. Finally, the coding stage is made with Wavelet Difference Reduction (WDR) with ROI priority which spends more bits in the interest zones to improve the reconstructed quality of the zones. A set of experiments demonstrate that this coder is more useful when bit rates is low.