The research and development of teleradiology and telemedicine systems has progressed through many technical and clinical endeavors. When dealing with large volume image transmission and storage, image data compression is an outstanding issue in medical applications to which current techniques were not designed to address. The technical objectives of this project are to develop optimized error-free as well as error-controllable methods for medical image compression based on wavelet transform and associated methods. In this project, we employ both advanced artificial intelligent and compression techniques to achieve these goals. Our recent research outcomes include: (a) development of a mathematics approach to unify prediction, subband, and wavelet transforms, (b) development of convolution neural network training methods to obtain optimized wavelet kernel, (c) development of a data splitting technique to improve edge accuracy and to provide error-control methods, and (d) development of an integer implementation method for all wavelet transforms, etc. Based on the above technical advances, we propose to use integer form of an adaptive (optimized) wavelets in conjunction with newly developed coding methods such as """"""""partitioning in hierarchical trees"""""""" (PHT) for lossless compression. For error-controllable approaches, we propose to use adaptive wavelets coupled with optimized neural network prediction methods in this study. Since lossless compression is a part of the error - controllable method, both systems can be implemented in the same scheme which is a breakthrough approach in the field. We will compare the compression results (i.e., compression ratio and speed) of the proposed compression methods with those of the current wavelet techniques using the embedded zero-tree coding method. At the end of the project, we will deliver a software package for the radiological society. Hence, the evaluation for various clinical applications using the proposed methods can be performed by the investigators. As the field of telemedicine is rapidly growing, we believe that development of a dedicated compression module for economical storage and fast communication of patient data (particularly for patient images) is necessary. This project is designed to address the related technical issues with a strong clinical consideration.

Agency
National Institute of Health (NIH)
Institute
National Cancer Institute (NCI)
Type
Research Project (R01)
Project #
1R01CA079139-01
Application #
2710010
Study Section
Biomedical Library and Informatics Review Committee (BLR)
Program Officer
Menkens, Anne E
Project Start
1998-04-07
Project End
2001-03-31
Budget Start
1998-04-07
Budget End
1999-03-31
Support Year
1
Fiscal Year
1998
Total Cost
Indirect Cost
Name
Georgetown University
Department
Radiation-Diagnostic/Oncology
Type
Schools of Medicine
DUNS #
049515844
City
Washington
State
DC
Country
United States
Zip Code
20057
Lo, Shih-Chung B; Li, Huai; Freedman, Matthew T (2003) Optimization of wavelet decomposition for image compression and feature preservation. IEEE Trans Med Imaging 22:1141-51