Learning and Diagnostics of Lymphocytic Thyroiditis based on Sonogram Scans

L. Tesař, D. Smutek, J. Jiskra, and P. Maruna (Czech Republic)


Pattern recognition, medical image analysis, Bayesian approach, Gaussian mixture modeling, Hashimoto's lymphocytic thyroiditis.


The problem of diagnostics of diseases of thyroid gland is addressed. We divide this problem into two stages. In learning stage, the database of sonograms with known diag nosis is used. In diagnostics stage, the sonogram of person with unknown diagnostics is taken and the most probable diagnosis is inferred. The paper uses Bayesian inference with Gaussian clusters of pattern features. The method is tested on Hashimoto's lymphocytic thyroiditis, which is ideally suited for this approach.

Important Links:

Go Back