Adaptive Pixel-Based Technique using MMSA for Grayscale Image Compression

This paper investigates APBT compression system for grayscale natural images to achieve a higher CR with preserving image quality by utilizing C321, DWT, BPS for MSV/LSV. The APBT efficiently improves the performances compared to TPBT, by overcoming the inherited problems related to modeling parts (deterministic /probabilistic), due to utilizing image data redundancy(s) successfully, where the deterministic part (Mn, dx) reduced by over 400 bytes, and probabilistic part (Rs) reduced by more than 5000 bytes on average. So the required size in bytes has been reduced by nearly 50% compared to TPBT, in which the CR increased twice the traditional one on average, with higher quality by 5 dB more compared to the TPBT.