Adaptive grayscale image coding scheme based on dynamic multi-grouping absolute moment block truncation coding
- PDF / 1,965,612 Bytes
- 17 Pages / 439.37 x 666.142 pts Page_size
- 70 Downloads / 252 Views
Adaptive grayscale image coding scheme based on dynamic multi-grouping absolute moment block truncation coding Jun-Chou Chuang 1 & Yu-Chen Hu 2 & Chia-Mei Chen 3 & Zhaoxia Yin 4 Received: 20 August 2019 / Revised: 30 June 2020 / Accepted: 9 July 2020 # Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract
Multi-Grouping Absolute Moment Block Truncation Coding (MGAMBTC) technique improve the image quality of Absolute Moment Block Truncation Coding (AMBTC) by adaptively dividing pixels in a block into groups according to block activity. This study found that the bit rate can be reduced further if there is some similarity among the blocks. Therefore, this study proposes a block prediction scheme that exploits the intra-block similarity of neighboring blocks. If a similar encoded block is found, its position code will be stored to encode the block; otherwise, it is encoded with the proposed improved MGAMBTC (iMGAMBTC). The proposed iMGAMBTC employs an entropy-based indicator generation mechanism to reduce the bit rate, and the evaluation demonstrates that the proposed scheme enhances the compression performance of AMBTC and MGAMBTC efficiently. Keywords Image compression . Absolute moment block truncation coding . Adaptive pixel grouping . Block prediction technique . Entropy coding
1 Introduction Delp and Mitchell proposed the Block Truncation Coding algorithm (BTC) [9, 11, 25] for grayscale image compression which divides the pixels of an image block into two groups and
* Yu-Chen Hu [email protected] Jun-Chou Chuang [email protected] Chia-Mei Chen [email protected] Zhaoxia Yin [email protected] Extended author information available on the last page of the article
Multimedia Tools and Applications
calculates the quantization level of each group recorded by a binary bit map. Therefore, an encoded image block consists of two quantization levels and a binary bit map. BTC is also called moment preserving block truncation coding (MPBTC), as it preserves the first and second moments of a block by the two quantization levels. Lema and Mitchell [25] proposed an improved version of the previous work with better image quality, Absolute Moment Block Truncation Coding (AMBTC), where their major difference is the way of quantization level generation. BTC-based image coding techniques had been proposed. Studies [5, 25] introduced frequently used bit map patterns to save storage space; Nasiopoulos et al. [28] proposed a bit map interpolation technique; Olsen [29] presented a fast image coding by planar block coding. An AMBTC-based image compression scheme [13] utilizes quadtree segmentation to exploit variable-sized block encoding. An improved MPBTC scheme [14] applies block search order coding to exploit the similarity of neighboring blocks. Dhara and Chanda [10] designed a pattern fitting scheme to find an optimum bit-pattern for binary bit map encoding. Hu proposed a predictive MPBTC for grayscale image compression [15], where a two-dimension prediction is employed to encode an image block if a
Data Loading...
