Texture Analysis of Thyroid Ultrasound Images for Diagnosis of Benign and Malignant Nodule using Scaled Conjugate Gradient Backpropagation Training Neural Network
Shrikant D. Kale, Krushil M. Punwatkar · 2013
A thyroid is largest endocrine gland, a butterfly shape with two lobes which produces hormones that control body metabolism. The nodules are found in thyroid may be benign or malignant. The ultrasound (US) preferred over the other medical imaging modalities which is used to observe the subcutaneous body structures & internal organs for possible pathology or lesions. The physicians are deducing useful information concerning the tissue characterization and structure which is still subjective matter, Thus quantitative analysis of US images give the objective method for thyroid nodule diagnosis. In this paper, gray level co-occurrence matrix (GLCM) texture characterization techniques are used for feature extraction & extracted features are classified using scaled conjugate gradient backpropagation training neural network (SCGBNN) for diagnosis of thyroid nodule are described. The experimental results show the performance measure of SCG backpropagation training neural network in terms classification accuracy.