Abstract
We propose a 2D VLSI architecture with 256 processing elements and a computation result reuse methodology for full search variable block size motion estimation (FSVBSME) in the next generation video coding standard H.265/AVC. Our pipelined engine can