A Wavelet Based Algorithm for Edge Extraction of Images

In this study, we proposed a contextual filter edge detector and a multiscale edge tracker to extract edges for further analysis of computer vision. The contextual filter detected significant edges in the finest scale gradient image. Many noised or blurred cases were suppressed with higher decision levels. The edge tracker further improves the detected results using the multiscale gradient images. Qualitative and quantitative comparisons with other methods show that the proposed approach extracts better edges than other detectors extract