In this paper, we propose a fast motion vector detection based on gradient method. We define objective function as squared error over a block. Proposed method obtains motion vector by computing the differential coefficient of the function and applying one dimentional search to the direction. By applying the hierarchical method, it can be avoided to fall in the local minima. From simulation experiments, we compare predictive performance and computational complexity with full search block-matching algorithm and show the effectiveness of the proposed method.
展开▼