Speaker:

Youn-Long Lin, National Tsing Hua University, Taiwan

Title:

Intra-Frame Video Compression for Memory Traffic and Memory Reduction

Abstract:

We propose a low-complexity intra-frame compression algorithm for video display frames. It combines a dictionary-based coding method, the Huffman coding method, and three innovative schemes to achieve high compression ratio. We quantitatively analyze the characteristics of display frame data for algorithm design. We first propose a two-level classification scheme to classify all pixels into three categories. Then we employ a dictionary-based coding method and propose an adaptive prefix bit truncation scheme to generate codewords for video pixels in each category. We subsequently employ the Huffman coding method to assign bit patterns to the codewords. Finally, we propose a head code compression scheme to further reduce the size of the codeword bits. Experimental results show that the proposed hybrid compression algorithm improves at least 10\% of data reduction ratio over prior arts. When simulated along with an H.264/AVC decoder, it also reduces by 59\% the number of bus cycles required to access display frames.

Bio:

Youn-Long Lin receives his PhD from the University of Illinois, Urbana-Champaign. He is a professor with National Tsing Hua University, Taiwan, since 1987. His research interest includes physical design automation, high-level synthesis and video coding. He is a co-founder of Global Unichip Corp.