Multiresolution edge detection in noisy images using wavelet transform
Qing-Hua Lu, Xianmin Zhang · 2005
The edge detection method in noisy images using multiresolution wavelet transform is presented in this paper. The noisy images are first decomposed by multiscale wavelet to different scales wavelet coefficients (horizontal, vertical, and diagonal direction), then the shifted wavelet coefficients are multiplied step by step to depress noise and detect edges. Compared to other methods, this approach not only can get rid of noise very well and decrease the number of computations, but also can detect diagonal edge information.