Cytological Detection of Thyroid Cancer by Optical Image Analysis

Abstract

This research explores the use of features of cells in digital optical images of human thyroid tissue as an important base to diagnose the cancer. It presents some efficient features of cells nuclei for detection of thyroid malignancy such as (radius, smoothness, compactness, expected value and variance). The cytological characteristics are very important and usual method to separate abnormal and normal cases in all diseases. The algorithm of neural network used to detect thyroid cancer successfully with accuracy of 99%.

Keywords: thyroid cancer, neural network, optical images, malignancy