Computer Imagery Analysis for Medical Applications
Recent medical image research has resulted in a new computer-aided diagnostic methodology called Computer-Aided Diagnostic Characterization (CADc). The method describes suspicious lesions using medically meaningful terminology to support radiologists in making diagnostic decisions based on computed tomography (CT) lung image scans.
Benefits of the new methodology include ratings that:
- Serve as measurements (direct evidence) of medically meaningful, visual signs of disease providing physicians with similar cases (in which the similarity is based on nodule appearance) and allowing automated case-based reasoning and prediction of the likelihood of disease based upon a database of known cases
- Allow users to retrieve visually similar cases, which can be useful for training new radiologists.
The CADc feature extraction and prediction approach is illustrated below.
For more information, contact:
Decision and Information Sciences Division
Argonne National Laboratory
9700 South Cass Ave., Bldg. 221
E-mail Craig Swietlik