Katar uzunluğu, zincir kod ve EZW kodlamaları tabanlı tıbbi görüntü sıkıştırma yaklaşımları

In information theory, utilization of channel bandwidth efficiently, establishing the practical telemedicine networks and even though archiving of medical images are importantly considered to transfer data. The expression of an image using a fewer number of bits with or without loss of information is defined as image compression or coding. In order to image coding, there are many compression techniques. In this dissertation, various compression algorithms are examined to reveal the redundancy of 3D medical images more effectively. The context-based and contour-based coding approaches are proposed bi-level compression pipelines. In the first circumstance, the run-length coding is specialized for two-dimensional slices of volumetric-medical images. Inter-voxel relationships and intra-pixel relationships are revealed by different scanning procedures such as Hilbert, chevron, and perimeter. Secondly, chain codes are applied to 2D-slices to code contour knowledge. In this method, the contour defining algorithm is used and modified to code symbols representation efficiently. The embedded zerotree wavelets (EZW) and sparsity are gray-level compression approaches, and they utilize different wavelets. In this study, proposed algorithms are experienced on the computed tomography (CT) and magnetic resonance imaging (MR) datasets, which are acquired from Dokuz Eylül University Hospital. The run-length and chain codes systems are applied for bi-level CT and MR datasets and compression ratios approximately 100:1 and 200:1, respectively. These achievements show proposed systems outperform JBIG and CCITT that are well-known bi-level compression standards. The EZW algorithm only tested for gray-level MR images. The results are presented in terms of common lossy compression metrics.