Multi-scale Based Adaptive SRAD for Ultrasound Images Enhancement

Abstract

These in this paper, we introduce the adaptive speckle reducing anisotropic diffusion(SRAD) that uses wavelet decomposition for speckle reduction and edge preservation in medical ultrasound images. The proposed algorithm first modifies coarse-to-fine classifier to determine the homogeneous region that operated as diffusion threshold in SRAD works. Next, SRAD filters are played on each sub-bands of decomposed wavelet domain with determined homogeneous regions as called ldquospeckle scale functionrdquo. From this procedure, homogenous region can calculated without manual selection or preliminary exponential decay function. As a result, original SRAD filter will modified as adaptive one. Moreover, variety pattern of speckle is reduced by the proposed modified SRAD on multi-scale wavelet decomposing. Relying on this progress, the proposed method can improve the image quality for both speckle reduction and edge preservation. Finally, we validate our method to compare with conventional filter groups using artificial speckle image and clinical images. The experimental results show that the proposed method performed effectively both terms of speckle reduction and edge preservation.