Medical image de-noising based on wavelet correlation thresholding

Edge-preserving de-noising of ultrasound and MRI images is of great interest in medical image applications. This paper presents a wavelet-based scale correlation thresholding scheme for medical image enhancement. The dyadic wavelet transform that acts like a Canny edge detector is employed to characterize the significant structures, which would be strongly correlated across the wavelet scales. A robust threshold is presented and the experiments on natural, MRI and ultrasound images show that the proposed scheme outperforms the traditional thresholding schemes 1701 only in SNR measurements but also in the edge preservation.