Investigation on fast motion search algorithms and area efficient deblocking filter for video coding standards

Abstract

Video coding is the technique to compress the original video to a smaller size without degrading the quality to a larger extent In video coding standards Motion Estimation ME is the major part to estimate the motion vector between the current frame and reference frame for coding The process of motion estimation utilizes 50 70 of total computational complexity For encoding the high resolution videos such as HD 720p and FULL HD 1080p videos the High Efficiency Video Coding HEVC standard is used The coding unit size of the HEVC is larger than 16×16 blocks which increases computational complexity of the encoder To reduce the complexity efficient motion estimation algorithms are proposed in HEVC In this thesis Enhanced Test Zone ETZ search motion estimation algorithm is proposed for HEVC which performs better than the existing Test Zone search algorithm with reduced computational complexity In this proposed algorithm a coding unit type based different predictor set and search methodologies with early termination are used To reduce the computational complexity with good Rate Distortion RD cost performance certain conditions are derived for early termination. newline