Space-variant preprocessing for transform-based medical image compression
Date of Completion
January 1990
Keywords
Engineering, Biomedical|Biophysics, General|Computer Science
Degree
Ph.D.
Abstract
Space-variant preprocessing combines the best of space and transform domain compression techniques so that energy packing efficiency can be higher than by conventional transform techniques. The fast Cas-Cas transform (FCCT) provides improved speed of computation as a transform method. A new symmetrical bit allocation table method is presented to compress efficiently all full frame transformed coefficients. The preprocessing approach combined with the Cas-Cas transform and the new symmetrical bit allocation table method yields improved performance over conventional methods when applied to modalities such as Magnetic Resonance Imaging(MRI), Ultrasound, and plain film X-ray. Overall space-variant preprocessing was found to significantly improve all transform based compression schemes tested, and provided around 10:1 compression ratios with no data loss other than quantization errors. ^
Recommended Citation
Paik, Chul Hwa, "Space-variant preprocessing for transform-based medical image compression" (1990). Doctoral Dissertations. AAI9109852.
https://digitalcommons.lib.uconn.edu/dissertations/AAI9109852