Motion estimation in ultrasound image using dynamic multi-shape search

Motion Estimation is a key problem in video processing and computer vision. It is the process of determining motion vectors that describe the transformation from one 2-Dimension or 3-Dimension image to another. Many algorithms for motion estimation have been proposed, but some are computationally expensive and some have quality limitations. Block-Matching is a widely-used method for stereo vision, Visual tracking, video compression, and objects motion estimation, in which images/videos extracted into macro-blocks, thereby achieving image/videos redundancy is reduced from one frame to another. In this paper a new novel approach, we call it Dynamic Multi Shape Search technique, has been proposed to efficiently estimate the location of atrial-septal-defect, which uses three steps with 21 points separated to 9 points in semi diamond shape, then 8 points with square shape and finally 4 points with diamond shape. The main feature of this approach is to allow a full search, while being accurate at the same time. In addition it takes less time than Four Step Search techniques. The contribution of this paper is to show that small or large motion cPan be estimated directly by using our novel Dynamic programming Multi-shape Search Technique instead the oldest block-matching computation algorithms. Our new approach was compared with baseline algorithms like three step Search and four step Search. The results showed the validity of our method as competitive results quality-wise, and a significant gain speed-wise.