Ultrasound B-scan image speckle noise reduction and boundary enhancement

Speckle problem is one of major issues in ultrasound B-Scan images. Most of proposed methods degrade the image features (boundaries, textures, etc.) while reducing the speckle effect. The proposed technique, minimum directional derivative search (MDDS) method, takes the image formation and coherent features into account to proceed the filtering. In a gray level field of an ultrasound B-Scan image, it can be shown that the image properties are least variant along the direction perpendicular to the gradient in a local point. Filtering along this direction won't mix different statistics and won't blur edges with their surroundings. A set of directional cancellation kernels has been designed to perform the minimum directional derivative search. Simply filtering the local data along this direction can receive significant result on speckle reduction and boundary enhancement compared to other filtering methods.