Thyroid Nodule Classification in Medical Ultrasound Images

摘要

In this paper, an algorithm for thyroid nodule assessment in medical ultrasound images is proposed. Ultrasound imaging is a noninvasive diagnostic tool to detect abnormalities of thyroid gland. But the presence of speckle noise limits the contrast resolution and makes diagnosis more difficult. Hence despeckling is used as a pre-processing step. The nodule region is segmented using Fuzzy C-Means clustering. Gray Level Co-occurrence Matrix (GLCM) features are extracted and are utilized for classification using Support Vector Machine (SVM). The performance evaluation of the SVM classifier is done using accuracy, sensitivity and specificity.