Computer Science · Computer Vision and Pattern Recognition
Adaptive anisotropic diffusion filter for speckle noise reduction for ultrasound images
Mohammad Motiur Rahman, Mithun Kumar PK, Abdul Aziz, Md. Gauhar Arefin, Mohammad Shorif Uddin
See on the mapAbstract
In image processing noise is possibly the most annoying problem. Ultrasound image has been used for effective diagnosis for its non-invasive, harmless, accurate and cost effectiveness. Unfortunately, ultrasound images are degraded by an intrinsic artefact called speckle. Eliminating such speckle noise is an important preprocessing task. For acquiring a better performance for denoising, an adaptive anisotropic diffusion technique for ultrasound images is presented in this paper. It is a new efficient method for denoising the image without blurring the frontiers between different regions. The proposed noise detection-oriented method is based on the differentiation of the diffusion in the direction of the gradient and the adjacent pixels of entire image. The performance of different metrics shows the proposed method exhibit better result to existing Perona-Malik anisotropic diffusion method.