Ultrasonic Thyroid Nodule Segmentation Based on Segmentation Adversarial Network

Abstract

An ultrasonic thyroid nodule segmentation model based on conditional segmentation adversarial network (cSegAN) was proposed to achieve more accurate segmentation of thyroid nodules. The model is composed of two parts: a segmenter network and a discriminator network. The segmenter network design uses a multi-expansion rate convolution block to accurately locate the nodule area, learn to extract the depth and shallow feature information of the nodule, and obtain binary mask of nodule area; the discriminator network compares the gap between the segmentation result and the gold standard to evaluate the segmentation result. Through multiple adversarial training, the experimental results show that the pixel accuracy of the proposed model reaches 0.953 1, which is better than other segmentation models, and can achieve ultrasonic thyroid nodule segmentation more accurately.

Keywords

• thyroid nodule segmentation
• convolutional neural network
• segmentation adversarial network
• ultrasound images
• adversarial training