(Taken from application abstract): Digital images obtained from diagnostic medical procedures and archived in high-speed networks are an immense scientific resource. Our laboratory has developed database techniques for organizing image archives in such a way that images can be indexed and retrieved on the basis of pictorially defined image features. We have developed a geometric reasoning engine which indexes images on the basis of """"""""explicit"""""""" and """"""""implicit"""""""" geometric properties of anatomy. We have evaluated the performance of our methods on a collection of static cardiac MRI images. The technique automatically organizes such a collection into image view-plane (axial, sagittal, oblique, etc.) without prior definition, and can structure image collections on a notion of """"""""shape"""""""" that correlates well with that of expert radiologists. We propose to extend these concepts to index moving image sequences of the type found in diagnostic cardiology. We will investigate the use of quantitative parameters derived from the analysis of non-rigid motion as a means to index image sequence data. We believe that we can apply our laboratory's experience with shape and non-rigid motion analysis to moving image sequences so that they can be ordered and retrieved on the basis of similarity of shapemotion, a concept which is useful in such disciplines as cardiac diagnostic function assessment. Although our approach is applicable to a number of imaging modalities, we chose the area of echocardiographic imaging as our focus in order to capitalize on our expertise and the availability of data. We have a large collection of digital echocardiographic video sequences gathered on CD-ROMs for educational purposes and we will index these with the techniques we propose to develop.

Agency
National Institute of Health (NIH)
Institute
National Library of Medicine (NLM)
Type
Research Project (R01)
Project #
5R01LM005007-09
Application #
2771691
Study Section
Biomedical Library and Informatics Review Committee (BLR)
Program Officer
Bean, Carol A
Project Start
1990-05-01
Project End
2000-08-31
Budget Start
1998-09-01
Budget End
1999-08-31
Support Year
9
Fiscal Year
1998
Total Cost
Indirect Cost
Name
Yale University
Department
Radiation-Diagnostic/Oncology
Type
Schools of Medicine
DUNS #
082359691
City
New Haven
State
CT
Country
United States
Zip Code
06520
Tagare, H D (1997) Deformable 2-D template matching using orthogonal curves. IEEE Trans Med Imaging 16:108-17
Robinson, G P; Tagare, H D; Duncan, J S et al. (1996) Medical image collection indexing: shape-based retrieval using KD-trees. Comput Med Imaging Graph 20:209-17