International Journal of Research in Arts and Science
ISSN: 2394-9759 | Frequency: 4 Issues/Year
Impact Factor: 0.387 | International Scientific Indexing(ISI) calculate based on International Citation Report(ICR)
Classification of Adenomatous Hyperplasia Thyroid Nodules Using the Features Extracted from Ultrasound Images
Abstract:
Ultrasound imaging technique of the thyroid gland is considered to be widely used diagnostic method for evaluating thyroid nodules. The ultrasound imaging is a radiation free, non-invasive technique and cost effective. Moreover, computer-aided diagnosis (CAD) are very helpful for radiologists, but with some limitations. One of the major limitations of CAD system is that it requires direction dependent features. In this study, classification of thyroid nodule as adenomatous hyperplasia using only the features extracted from ultrasound images is suggested. The intension is to design a CAD system that will use only direction independent features. Also, any feature extraction techniques such as Gray Level Run Length Matrix (GLRLM) and Gray Level Co-occurrence Matrix (GLCM) can be used to identify the features. The features may be then used in classifiers such as Support Vector Machine (SVM) to categorize the thyroid nodule as adenomatous hyperplasia of thyroid or normal class. The classification results are evaluated with histopathology results. This CAD system will help the radiologists for diagnosis of thyroid nodules.
Keywords: Adenomatous Hyperplasia, Thyroid Cancer, Thyroid Nodules, GLCM, GLRLM
Volume: 5 | Issue: Holistic Research Perspectives [Volume 4]
Pages: 208-218
Issue Date: August , 2019
 |
|