• Wrushali M. Mendre
• Mousami V. Munot

The automated detection of thyroid nodule malignancy is a thrust area of research. Literature reports popular use of Fine Needle Aspiration Biopsy [FNAB] to confirm the nodule malignancy, however there is a dreadful need to improve the ability of predicting thyroid malignancy in USG images prior to FNAB. The proposed research aims to design and develop a computerized system to aid the decision making about a malignant thyroid. This study exploits Discrete Wavelet Transform [DWT] to devise a feature vector with seven textural features and additionally fifteen features are derived using Gray Level Co-occurrence Matrix [GLCM]. A Support Vector Machine (SVM) is used as an intelligent classifier to identify the malignant thyroid nodules using the fusion of the derived features. Standard database from Wilmington Endocrinology comprising of 79 ultrasound images of thyroid nodule (51 benign and 28 malignant cases) are considered to examine the efficacy of the proposed algorithm. The performance of the SVM classifier is evaluated considering polynomial, Radial Basis Function [RBF], linear and Multilayer Perceptron [MLP] kernel functions and analyzing the respective Receiver Operating Characteristics [ROC]. Highest classification accuracy of 99.98 with Area under ROC Curve (AUC) of 0.995, employing the polynomial kernel of 3rd degree confirms the superiority of the proposed approach.

Ultrasound, Textural Features, GLCM, DWT

"Texture and Wavelet Based Classification of Thyroid Nodule in Ultrasound Image Using Kernel Based Support Vector Machines", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.6, Issue 6, page no.660-668, June-2019, Available :http://www.jetir.org/papers/JETIR1906872.pdf

"Texture and Wavelet Based Classification of Thyroid Nodule in Ultrasound Image Using Kernel Based Support Vector Machines", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.6, Issue 6, page no. pp660-668, June-2019, Available at : http://www.jetir.org/papers/JETIR1906872.pdf