Boosted Nonlinear Coherent Diffusion for Despeckling of Diagnostic Images

Ultrasonography is a powerful technique for imaging the internal anatomy with its nature of low cost, portability, non invasive and real time imaging formation compared with other imaging modalities. Speckle noise is an inherent nature of ultrasound images, which may have negative effect on image interpretation and diagnostic tasks. It is necessary to preprocess imagery to reduce granular, texture like noise called speckle. This preprocessing is difficult when it is needed to preserve delicate image details that are buried in speckle. In this paper we present an effective approach for speckle reduction in medical diagnostic ultrasound images. This method utilizes non linear coherent diffusion model with median regularization term as boosting source for the point and linear features in image being processed. Windowed second moment tensor is used as a diffusion tensor to estimate the local coherence. Experiments have been performed on synthetic image, simulated phantom and clinical ultrasound images. The results are compared with existing state of art methods and our proposed method gives comparatively better result in term of FOM, SSIM and EPI.