Performance Assessment of Various Thyroid Image Segmentation Techniques with Consistency Verification

Neetu Chikyal, K. Veera Swamy

It is an essential job of the physician to detect and identify the various causes of the dieses related to the Thyroid gland. One of the cheapest techniques to examine the thyroid gland is the ultrasound imaging technique which is free from ionizing radiation and have low risk. In this work various segmentation techniques like threshold, active contour and k-mean clustering are used for thyroid gland segmentation. Segmentation consistency verification is applied to improve the activity. Performance of various segmentation algorithms are assessed using accuracy, sensitivity, specificity, positive and negative prediction value. Experimental results indicate that consistency verification gives better results.

Volume 11 | 02-Special Issue

Pages: 1299-1309