Journal article
Teaching the Computer Pattern Recognition: Modeling the Visual System
Proceedings - Symposium on Computer Application in Medical Care, pp.751-755
08 Nov 1989
Abstract
• One can teach the computer to recognize nerve cells by modeling algorithms after the human visual system. The program described here, DIVAR (Database Independent Visual Algorithms of Recognition) employs feature-detection units that have the ability to communicate their outputs amongst themselves, simulating the brain's parallel processing architecture. The program has a sensitivity of 84% with 4% error. These results are realized because recognition is not based on definitive decisions at each level of processing; rather, it is part of a statistical process in which the units settle to a solution using functional and spatial parallelism, alternating between looking at the parts of the pattern and the pattern as a whole. In time, artificial intelligence, including patern recognition, will be an invaluable aid in research and medical diagnosis.
Metrics
1 Record Views
Details
- Title
- Teaching the Computer Pattern Recognition: Modeling the Visual System
- Creators
- Jonathan KraidinArnold J. SmolenPatricia Beaston-Wimmer
- Publication Details
- Proceedings - Symposium on Computer Application in Medical Care, pp.751-755
- Number of pages
- 5
- Resource Type
- Journal article
- Language
- English
- Academic Unit
- MD (Doctor of Medicine) Program
- Identifiers
- 991021961108204721