Predictor Based Block Adaptive Near-Lossless Coding Technique for Magnetic Resonance Image Sequence

In today's world of information exchange, digital imaging of pathological information has grown very fast and hence plays an essential role in disease diagnosis. Radiological modalities such as Computed Tomography (CT) scanners, Magnetic Resonance Imaging (MRI), X-Ray are increasing massively. The demand for handling images in digital form has increased dramatically in recent years for efficient archiving and transmission that raises the need of efficient compression. In this paper, a near-lossless predictive based coding technique is proposed that address the low coding efficiency of lossless coding and low image quality of lossy coding. The proposed technique removes inter-pixel redundancy and physco-visual redundancy by employing Resolution Independent Gradient Edge Detector (RIGED) at optimal threshold value and optimal q-level for quantization respectively. Coding redundancy is removed by employing block based encoding at optimal fixed size block. Volumetric MR image datasets of 8 bit and 16 bit are utilized for implementation of the proposed technique. The proposed technique is compared with the state-of-the art techniques and outperforms Context Adaptive Lossless Image Compression (CALIC), Low Complexity Lossless Compression for Images (LOCO-I) and Differential Pulse Code Modulation (DPCM) with context switching NN predictor by 34.33%, 39.80% and 30.35% respectively. (C) 2020 The Authors. Published by Elsevier B.V.