Dictionary Approaches to Image Compression and Reconstruction
Nigel Ziyad, E. Thomas Gilmore, Mohamed Chouikha · 1998
This paper proposes using a collection of parameterized waveforms, known as a dictionary, for the purpose of medical image compression. These waveforms, denoted as f g , are discrete time signals, where g represents the dictionary index. A dictionary with a collection of these waveforms is typically complete or overcomplete. Given such a dictionary, the goal is to obtain a representation image based on the dictionary. We examine the effectiveness of applying Basis Pursuit (BP), Best Orthogonal Basis (BOB), Matching Pursuits (MP), and the Method of Frames (MOF) methods for the compression of digitized radiological images with a wavelet-packet dictionary. The performance of these algorithms is studied for medical images with and without additive noise. Keywords: Image coding and compression, Medical Image Processing, Signal Reconstruction, Wavelets Prepared through collaborative participation in the Advanced Telecommunications/Information Distribution Research Program (ATIRP) Consortiu...