An Adaptive Deblocking Algorithm with Boundary Detections

• Main Authors: Yi-ling Chen, 陳怡伶
• Other Authors: Shen-chuan Tai
• Format: Others
• Language: en_US
• Published: 2008
• Online Access: http://ndltd.ncl.edu.tw/handle/38661370093235616373

碩士 === 國立成功大學 === 電腦與通信工程研究所 === 96 === A new method to detect block boundaries and remove blocking artifacts in low bit-rate block-based images and video sequences is proposed in this thesis. The detection stage is applied to detect the location of block boundaries in deblocking frames, including measuring the width and the height of blocks and locating where the boundaries are. The deblocking algorithm is applied to the block boundaries which are detected at the detection stage. A low-complexity deblocking filter with three modes is proposed, including smooth mode (for low-frequency regions), complex mode (for high-frequency regions), and corner mode (for corner of four adjacent blocks). The mode decision procedure is based on observing pixel behavior around the block boundaries. The proposed algorithm can detect the block boundary with high accuracy, and remove more blocking artifacts.