Title: Model-based image compression framework for CT and MRI images

Authors: Sandeep Jain, Ankush Mittal, Sujoy Roy

Addresses: Department of Computer Science and Information Technology, Jaypee Institute of Information Technology, Noida (U.P.) – 201307, India. ' Computer Science and Engineering, College of Engineering Roorkee, Uttarakhand – 247667, India. ' Computer Vision and Image Understanding, Institute for Infocomm Research, 138632, Singapore

Abstract: This paper presents a novel model-based compression method for medical images. Unlike existing model-based compression methods, the proposed method is truly lossless and provides good compression ratio. The proposed method entails two major operations: registration of model and input (to be compressed) images, and compression of residual image (difference of model and input image). A context-based registration technique is proposed that depends on the criterion of minimising the residual image size. A quadtree-based adaptive block partitioning with rearrangement (QABPR) compression scheme is used to compress the residual image. Comparison with existing medical image compression methods and standard lossless compression techniques shows promising results.

Keywords: model-based compression; residual compression; context-based registration; medical informatics; block partitioning; image compression; CT images; MRI images; computed tomography; magnetic resonance imaging; medical images.

DOI: 10.1504/IJMEI.2011.039075

International Journal of Medical Engineering and Informatics, 2011 Vol.3 No.1, pp.40 - 52

Published online: 28 Feb 2015 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article