Related Experiment Videos
Artificial neural networks and their use in quantitative pathology
1Department of Obstetrics and Gynecology, University of Chicago, Illinois 60637.
Analytical and Quantitative Cytology and Histology
|December 1, 1990
Summary
Artificial neural networks (ANNs) offer a novel approach to quantitative pathology by learning from data to solve pattern recognition problems. ANNs demonstrate potential as valuable software tools when combined with other AI techniques.
Area of Science:
- Computer Science
- Biomedical Engineering
- Pathology
Background:
- Artificial neural networks (ANNs) are computational models inspired by biological neural networks.
- Quantitative pathology involves the objective measurement and analysis of biological samples.
- Traditional methods may struggle with complex, noisy, or ambiguous data in pathology.
Purpose of the Study:
- To introduce artificial neural networks (ANNs) and their structure.
- To demonstrate the application of ANNs in quantitative pathology for pattern recognition.
- To compare ANNs with traditional symbolic expert systems.
Main Methods:
- Detailed examination of a prototype neural network for pattern recognition.
- Exploration of neurocomputer technology applications in pathology.
- Analysis of ANN characteristics like data tolerance and generalization.
Main Results:
- A prototype neural network successfully addressed a pattern recognition task in quantitative pathology.
- ANNs showed capability in handling ambiguous, noisy data and generalizing from examples.
- Examples illustrated ANN use in pattern recognition, database analysis, and machine vision.
Conclusions:
- Artificial neural networks (ANNs) show promise for quantitative pathology applications.
- ANNs, combined with other AI and algorithmic methods, can serve as effective software engineering tools.
- The study highlights the strengths of connectionist approaches over purely symbolic systems in specific contexts.