Direction-adaptive fixed length discrete cosine transform framework for efficient H.264/AVC video coding

作者：

Highlights：

• Direction-adaptive fixed length-DCT(DAFL-DCT) is proposed for directional video data.

• Two versions of DAFL-DCT are suggested i) DAFL-HE and ii) DAFL-LC based on complexity.

• The best suitable directional transform mode is selected for 4x4 and 8x8 blocks.

• A new zigzag scanning pattern is proposed for 1D-DAFL-DCTs in inter-frame coding.

• To minimize transform modes overhead bits, side information encoding is proposed.

摘要

Highlights•Direction-adaptive fixed length-DCT(DAFL-DCT) is proposed for directional video data.•Two versions of DAFL-DCT are suggested i) DAFL-HE and ii) DAFL-LC based on complexity.•The best suitable directional transform mode is selected for 4x4 and 8x8 blocks.•A new zigzag scanning pattern is proposed for 1D-DAFL-DCTs in inter-frame coding.•To minimize transform modes overhead bits, side information encoding is proposed.

论文关键词：Discrete cosine transform,Direction adaptive transform,Directional transform framework,Video coding,Video compression

论文评审过程：Received 10 October 2015, Revised 2 September 2016, Accepted 4 September 2016, Available online 6 September 2016, Version of Record 10 September 2016.

论文官网地址：https://doi.org/10.1016/j.image.2016.09.002