Automatic classification of thyroid histopathology images using multi-classifier system

A computer aided diagnosis system supports doctors by providing quantitative diagnostic clues from medical data. In this paper, we propose a computer aided diagnosis (CAD) system to automatically discriminate hematoxylin and eosin (H&E)-stained thyroid histopathology images either as normal thyroid (NT) images or as papillary thyroid carcinoma (PTC) images. The CAD system incorporates a multi-classifier system to maximize the diagnostic accuracy of classification. Thyroid histopathology images are provided as input to the CAD system. The input images are enhanced and the nuclei present in the images are segmented automatically. Shape and texture features are extracted from the segmented images. Classification of the features is studied using classifiers such as support vector machine (SVM), naive Bayes (NB), K-nearest neighbor (K-nn) and closest matching rule (CMR) either as stand alone classifiers or as combinations to form multi-classifier systems. The multi-classifier system which provides the best accuracy is found out experimentally. The CAD system thus formed can be used as a second opinion to assist pathologists.