Thyroid Cancer Detection Using Thermal Images

1J.V. Vidhya, Amirtha Dasarathi and J.S. Lavannya


Thyroid cancer falls under the category of endocrine carcinomas. For many years, ultrasonography has been in use for the detection of thyroid cancer because of its good distinction between benign nodules and malignant nodules. Also, due to its better revealing of the pathological features, it has been preferred over CT and MRI scans. Despite this, the ultrasonography is not a very reliable method because of its dependency on Operator. With emerging trends, came the existence of Computer aided diagnosis (CAD). This method depends variably on the operator for resolving the subjective diagnosing problem. The objective of this project is to improve the accuracy of detection using various Image processing techniques, so that the tumor could be detected in early stage, preventing the delay in treatment and loss of life of the patient. The dataset that would be used for this process is Thermal images. Radiation in the long – infrared range of electromagnetic spectrum are detected using thermographic cameras and the resultant images are known as thermograms. The images would undergo pre-processing, segmentation and feature extraction to refine the input image, so that the tumor could be easily identified by the machine. After refining of the images, they would be subjected to classification process using Support Vector Machine to classify the input image as a malign cancer or benign cancer. This would result in determining the stage of cancer and thus eventually help in the further treatment process.


Thyroid Cancer, Thermograms, Support Vector Machine, Median Filter, Feature Extraction.

Paper Details
IssueIssue 5