Speckle Suppressing Anisotropic Diffusion Filter for Medical Ultrasound Images

Abstract

Ultrasonography is often preferred over the other medical imaging modalities due to its noninvasive nature, cost-effectiveness, and portability. However, the resolution of the ultrasound image greatly depends upon the presence of speckle noise. Speckle noise generally tends to reduce the image resolution and contrast, thereby reducing the diagnostic resolution of this imaging modality. In this paper, we propose a speckle suppressing anisotropic diffusion (SSAD) filter, to remove the speckle noise from B-Mode Ultrasound images. The performance of the SSAD filter is compared with the existing diffusion filters. The evaluation is based on their application to images simulated by Field II (developed by Jensen et al.). The algorithms were also tested for clinical ultrasound images of polycystic ovaries obtained from HDI 5000 Ultrasound Scanner. Performance evaluation was done by both numerical and functional parameters. The proposed filter yields better results in terms of greatest structural similarity index map (SSIM) of 0.95 and accuracy of 99.5.