Related Experiment Videos
Knowledge-guided segmentation of colorectal histopathologic imagery
D Thompson1, P H Bartels, H G Bartels
1Optical Sciences Center, University of Arizona, Tucson 85721.
Analytical and Quantitative Cytology and Histology
|August 1, 1993
Summary
A knowledge-based system successfully segmented colorectal gland images, achieving 85% accuracy. This method aids in analyzing histologic sections for conditions like adenomatous glands.
Area of Science:
- Histopathology
- Medical Image Analysis
- Computational Pathology
Background:
- Histologic sections of colorectal glands present segmentation challenges due to staining.
- Accurate identification of histologic components is crucial for disease analysis.
Purpose of the Study:
- To develop and evaluate a knowledge-based system for scene segmentation of colorectal gland images.
- To improve the accuracy of identifying and reconstructing histologic components in challenging image data.
Main Methods:
- A knowledge-based system was constructed for image segmentation.
- Processing sequences and a crowding index for adenomatous glands were computed.
- The system used context and logic to classify detected object groups within histologic sections.
Main Results:
- The knowledge-based system achieved acceptable segmentation in approximately 85% of processed colorectal gland images.
- The system successfully differentiated between various histologic components despite segmentation difficulties.
Conclusions:
- Knowledge-based systems offer a viable approach for accurate scene segmentation in complex histopathology images.
- This method facilitates the reconstruction and analysis of colorectal gland structures, aiding in diagnostic processes.