A new fast full search block matching algorithm using frequency domain

Abstract

The overwhelming complexity of block matching algorithm using a full search has prompted many companies and academic researchers to propose a myriad of algorithms. The challenge is to decrease the computational complexity of the full search as much as possible without losing too much performance and quality at the output. In this paper, we propose a new and fast algorithm which achieves exactly the same optimal result as the direct full search algorithm. The key idea is to express a robust matching criteria sum square difference (SSD) in terms of cross correlation operations. Speed is obtained from computing the cross correlations in the frequency domain via the Fast Fourier Transform (FFT). We also present a comparative performance analysis, which shows that the proposed method greatly outperforms the state-ofthe-art in frequency-domain motion estimation.