Classification of Microcalcifications for the Diagnosis of Breast Cancer Using Artificial Neural Networks.
Abstract
Early detection of breast cancer depends on the accurate classification of microcalcifications. We have developed a computer vision system that can classify microcalcifications objectively and consistently to aid radiologists in the diagnosis of breast cancer. A convolution neural network (CNN) was employed to classify benign and malignant microcalcifications in the radiographs of pathological specimen that were digitized at a high resolution of 21 microns x 21 microns. The CNN achieved an Az value of 0.90 in classifying clusters of microcalcifications associated with benign and malignant processes. An automated image feature extraction technique and feature-based neural network optimized with generic algorithms were applied to clinical mammograms as an alternative approach to the classification. The neural network system performed better than a radiologist in distinguishing between benign and subtle malignant clusters. We also developed an image display and analysis system that allows interactive 3D image manipulation and qualitative analysis of selected image regions of Breast MRI. This computer visualization system can help radiologists improve the efficacy of examining the massive amount of data, making BMRI a cost effect procedure with high sensitivity and specificity in the diagnosis of breast cancer.
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 01, 1997
- Accession Number
- ADA341532
Entities
People
- Yuzheng C. Wu
Organizations
- Georgetown University