IEEE Transactions on Pattern Analysis and Machine Intelligence
Pattern Classification (2nd Edition)
Pattern Classification (2nd Edition)
On reproducibility of ultrasound image classification
IbPRIA'05 Proceedings of the Second Iberian conference on Pattern Recognition and Image Analysis - Volume Part II
| Hi-index | 0.00 |
Ultrasonography is a cheap and quick non-invasive medical imaging technique, used as a diagnostic method for autoimmune thyroiditis. Another important diagnostic method for this chronic inflammation is measuring the increased level of thyroid antibodies (anti-thyroid peroxidase (TPOAb) and/or anti-thyroglobulin (TgAb)) in blood samples. This paper shows that B-mode ultrasound images contain weak information related to the presence or absence of these antibodies. Ultrasound image analysis is based on textural recognition using probabilistic spatial features. Two studies are performed. The results of the first study show that the spatial texture features we used contain weak information about the presence or absence of TPOAb and TgAb antibodies as measured by conditional entropy. In the second study, a classifier derived from Bayesian decision theory is tested on a set of 2820 sonograms of 94 subjects. The training set contains 67 subjects and the test set consists of 27 independent subjects. The results of classification to three classes (healthy, thyroiditis with positive antibody test, thyroiditis with negative antibody test) achieved sensitivity 29% and specificity 100% on the test set.