A Hardware-Oriented Image Compression Algorithm Based on BTC and YEF Color Space

• Main Authors: Ke, Shih-Yao, Jhou, He-Sheng, Chen, Chiung-An, Lin, Ting-Lan, Abu, Patricia Angela R, Chen, Shih-Lun
• Format: text
• Published: Archīum Ateneo 2021
• Subjects: training; image coding; quantization (signal); image color analysis; transform coding; prediction methods; prediction algorithms; YEF color space; block truncation coding; bit map; threshold; image compression; color sampling; Huffman coding; Computer Sciences; Databases and Information Systems
• Online Access: https://archium.ateneo.edu/discs-faculty-pubs/220; https://ieeexplore.ieee.org/document/9427696
• Institution: Ateneo De Manila University

This paper proposes a new compression algorithm based on color sampling and Block Truncation Coding (BTC). It is composed of several steps; color sampling; BTC parameter training; threshold selection; sub-sampling; prediction and quantization; and Huffman coding. The color sampling was applied using YEF color space which reduced 34% storage. The subsequent sub-sampling; quantization table prediction; and Huffman coding further increase compression rates. In addition; the BTC parameter training step finds the best reconstruction value for each 4*4 block and bitmap threshold value to improve the image quality. Compared with previously studies; the proposed algorithm shows a better average compression rate.