A new embedded hardware architecture for global motion estimation in MPEG-4

Global motion estimation (GME) is one of the key modules in MPEG-4 advanced simple profile (ASP). In this paper, a hardware-oriented GME algorithm is proposed, which is a combination of feature-points extraction and tracking algorithms using gradient methods. Our approach improves existing state-of-the-art GME algorithms by introducing two major modifications: first, only a small subset of the original image pixels is used in the estimation process. Second, a new associated parallel hardware architecture is proposed. It combines a content addressable memory (CAM) and digital signal processor chips. The experimental results show that the performance of the proposed hardware architecture can achieve the requirements of MPEG-4 with 4 CAMs and 1 DSP at 40 MHz working frequency.