Near-lossless image compression algorithm to reduce the memory bandwidth of display devices

초록

Abstract: Recently, as the spread of ultra-high definition (UHD) display devices has become more active, considerable data bandwidth is required between the display and external memory. Compression of data can be applied to reduce memory bandwidth, but that can lead to degradation of image quality. So, in this paper, we propose a near-lossless compression algorithm that reduces memory bandwidth and minimizes image quality degradation for hardware implementation. The proposed algorithm is based on two-line memory based prediction and transformation. Also, it improves the performance of quantization through one-dimensional (1-D) discrete cosine transform (DCT). In rate control, quantization parameters are determined using both horizontal and vertical information to improve efficiency. Finally, a partial update is implemented to enable selective decoding of images. Experimental results show that the proposed method is superior in terms of peak signal-to-noise ratio while maintaining a high compression ratio. © 2020 Institute of Electronics and Information Engineers. All rights reserved.

키워드