Thyroid myeloid cancer ultrasonic image recognition method based on clinical prior knowledge guidance

Abstract

The invention relates to a thyroid myeloid cancer ultrasonic image identification method based on clinical prior knowledge guidance. Firstly, a thyroid nodule area is obtained by designing a cascade segmentation network, then a multi-branch classification network is designed, and according to clinical knowledge guidance of radiologists, important thyroid ultrasound image feature validity and calcification are added as prior knowledge to assist network classification. The method not only can distinguish benign and malignant nodules in the image, but also can accurately distinguish the malignant nodules into myeloid thyroid carcinoma and papillary thyroid carcinoma.

Description

technical field [0001] The invention belongs to the field of image segmentation and classification, and in particular relates to an ultrasound image recognition method for medullary thyroid carcinoma guided by clinical prior knowledge. Background technique [0002] Most of the existing thyroid ultrasound image-aided diagnosis systems are for the classification of benign and malignant nodules, especially for the classification of thyroid papillary carcinoma. In contrast, medullary thyroid carcinoma has features of both benign and malignant nodules on ultrasound images (eg, figure 1 shown), it is difficult to accurately diagnose medullary thyroid carcinoma from images. At the same time, due to the low quality of ultrasound images and the existence of artificial artifacts, in order to obtain a better model, CNN often needs to train a large number of pictures, and the characteristics of medullary thyroid cancer cases are small, so it is relatively difficult to obtain so many pi...

AI Technical Summary

Problems solved by technology

[0008] 1. Medullary thyroid carcinoma only accounts for 2-3% of all thyroid cancers, and its data are scarce, and medullary thyroid carcinoma has both benign and malignant nodules in ultrasound images, so it is difficult to accurately diagnose medullary thyroid from images Therefore, it is difficult to construct an ultrasound image data set of medullary thyroid carcinoma, which makes the existing thyroid ultrasound image-aided diagnosis system not capable of identifying medullary thyroid carcinoma.

[0009] 2. The existing classification methods usually only consider some common features of the image when classifying thyroid nodules, and do not take into account the important discriminant features (such as calcification and solid part features) in thyroid ultrasound images

Images