A VLSI design of hierarchical search motion estimation processor chip

Abstract

This paper presents a motion estimation processor that has regular and simple structure and achieves 100% hardware utilization without image data fill time. It can compute half-pel precision estimation and I/O bottleneck is eliminated using a small distributed on-chip image memory. The number of processing elements is scalable according to the degree of parallel processing and throughput requirements. It has been designed and verified with C++ and VHDL.