Background
This research is part of the NCI Guanacaste project for studying visual features correlated to the development of precancerous lesions. The most important observation in a cervigram image is the Acetowhite (AW) region, which is caused by whitening of potentially malignant regions of the cervix epithelium, following application of acetic acid to the cervix surface. The image database consists of a very large archive of 60,000 digitized uterine cervix images, created by the National Library of Medicine (NLM) and the National Cancer Institute (NCI). These images are optical cervigram images acquired by Cervicography using specially designed cameras for visual screening of the cervix.
Methodology
We propose the use of powerful machine learning techniques such as Support Vector Machines (SVM) to learn, from a database with ground truth annotations, critical visual signs that correlate with important tissue types and to use the learned classifier for tissue segmentation in unseen images. In our experiments, SVM performs better than un-supervised methods such as Gaussian Mixture clustering, but it does not scale very well to large training sets and does not always guarantee improved performance given more training data. To address this problem, we combine SVM and clustering so that the features we extracted for classification are features of clusters returned by the mean-shift clustering algorithm. Compared to classification using individual pixel features, classification by cluster features greatly reduces the dimensionality of the problem, thus it is more efficient while producing results with comparable accuracy. (See more)
Multi-observer segmentation evaluation is useful in the imaging community. We have developed web-based software for automatic performance evaluation of multiple image segmentations which is based on the Baysian Decision framework. It computes a probabilistic estimate of the true segmentation(ground truth map) and performance measures for the individual segmentations (sensitivity and specificity). The strength of the tool is that it integrates the two kinds of prior knowledge of segmentations: the truth prior (the prior probability) and the observer prior (the performance measures of observers), which can generate more accurate evaluations. This tool has been used for cervigram images. (See more)
Publications
X. Huang, W. Wang, Z. Xue, S. Antani, L. R. Long, J. Jeronimo, "Tissue Classification using Cluster Features for Lesion Detection in Digital Cervigrams", SPIE Medical Imaging, San Diego, 2008
Y. Zhu, W. Wang, X. Huang, D. Lopresti, R. Long, S. Antani, Z. Xue, G. Thoma, accepted for publication in Journal of Signal Processing Systems: Special Issue on Biomedical Imaging. 2008, to appear.
Y. Zhu, X. Huang, D. Lopresti, L.R. Long, S. Antani, Z. Xue, G. Thoma, "Web-based Multi-observer Segmentation Evaluation Tool", to appear in 21st IEEE International Symposium on Computer-Based Medical Systems, Jyvaskyla, Finland, 2008
Contacts
