MR image compression with ROI decoding, SNR and Spatial scalability
Ramesh G. Naidu, V K Agrawal, AG Ramakrishnan · NOT FOUND REPOSITORY (Indian Institute of Science Bangalore) · 2008
This paper presents an algorithm that performs lossless compression of 3D magnetic resonance (MR) images and is capable of performing user selectable region of interest decoding together with spatial and signal-to-noise ratio (SNR) scalability. Selective decoding is achieved by segmenting the 3D MR image into blocks of size 16x16x2 in spatial domain and by separately coding each of the blocks. Spatial and SNR scalability are achieved by scanning of the wavelet coefficients in proper order, together with storing the markers for the bit stream generated. The encoding operation is performed only once for a given image; the quality parameters, such as spatial resolution, SNR resolution and the region of interest, can be defined at the time of decoding.