Motion Estimation and Compensation in Frequency Domain

摘要

In this paper, a novel method for motion estimation and compensation in frequency domain for the purpose of video compression is proposed. The proposed scheme is based on the coefficients of frequency domain obtained from the image blocks after the application of Orthogonal Polynomials Transformation (OPT). It uses the three-step search algorithm to estimate the search area for finding the best match in the subsequent frame with respect to each block in the current frame. The quality of match found will be contingent upon the value of Mean Absolute Difference, more commonly known as MAD, between the blocks. The motion vectors produced during motion estimation are utilized in the motion compensation process in order to produce the predicted frame for subsequent compression. The performance of the proposed scheme is experimented with good number of standard video and measured with Mean Square Error (MSE). The results are encouraging.