Adaptive two-step adjustable partial distortion search algorithm for motion estimation

Abstract

Motion estimation is widely used in video coding schemes, because it enables the transmission and storage of video signals within lower bit rate. A full search (FS) algorithm is optimal method of motion estimation, but it is suffers from high computational complexity. To reduce the complexity various methods are proposed. Recently, two-step edge based partial distortion search algorithm (TS-EPDS) is introduced, and it shows about 100 times faster than FS without PSNR degradation. In this paper, we proposed adaptive two-step adjustable partial distortion search algorithm, and it is 200 times faster than FS with negligible PSNR decrease. The proposed algorithm is suitable for real time implementation of high quality digital video application.