A parallel processing architecture for FSS block-matching motion estimation

Abstract

Motion Estimation (ME) is a key factor for achieving enhanced compression ratio. However, ME involves high computational complexity. The main goal is to reduce the execution time without reducing image quality. In this paper, a proposed high parallel processing architecture is presented for four-step search block-matching motion estimation. The proposed method develops an architecture which using 9 processing-elements (PE) and processes them in parallel. The architecture has been simulated and synthesized with VHDL and ASIC (CMOS 45nm). Synthesize results show that the proposed architecture achieves a high performance for real time motion estimation.