Improved k-d tree-segmented block truncation coding for color image compression
Ryan Rey M. Daga · 2017
Transmitting and storing digital images have bandwidth and disk space requirements. Reducing the file size of these images enables faster transmission of data and increases the number of images that can be stored in the same amount of disk space. Block truncation coding (BTC), one class of compression technique, is commonly used for its low computational complexity which make it suitable for multiple applications. A recently proposed compression technique, referred to as k-d Tree-Segmented Block Truncation Coding (KTS-BTC), was able to reduce the bit rate of the compressed image while maintaining image quality. In this study, we propose to improve KTS-BTC by implementing modifications: (1) implementation of Huffman Coding, and (2) encoding RGB values using shaved bit strings representing numbers that are divisible by a pre-defined power of 2.