Image Data Compression and Decompression Unit Considering Brightness Sensitivity

To reduce the large amount of memory required in image processing, high image quality memory data compression/decompression algorithms are proposed to eliminate the upper bits that overlap between neighboring pixels by the addition of simple hardware without image quality degradation inherent in conventional lossy compression. In addition, a method to maximize compression rates is discussed based on look-up tables, considering the human visual system. According to the characteristics of the frame image, a programmable architecture is proposed to compute proposed compression/decompression methods as well as run-length encoding and also conventional lossless algorithms. Also, compression rates and performances for SOP implementation are evaluated with the design and simulation of the proposed compression memory circuits.