Ultrasonic image thyroid nodule classification method based on feature decoupling

Abstract

The invention discloses an ultrasonic image thyroid nodule classification method based on feature decoupling, which is applied to the technical field of image processing and aims at solving the problem that the prior art is limited to change the input of a neural network to obtain features under different views. According to the method, a new local / global feature extraction method is established, a tissue-anatomy decoupling module is designed by using a self-attention mechanism to connect two paths of What and Where, and the task of thyroid nodule benign and malignant classification is completed in a multi-task learning form; according to the method, local / global feature decoupling can be completed in a feature space in a self-adaptive mode, compared with an existing method, a larger view field range is achieved, and more effective and stable features can be obtained. The benign and malignant thyroid nodule classification under the ultrasonic image has better diagnosis performance than doctors.

Description

technical field [0001] The invention belongs to the technical field of image processing, and in particular relates to a feature decoupling-based thyroid nodule classification technology in ultrasonic images. Background technique [0002] Thyroid nodules are common nodular lesions. According to the epidemiological investigation of thyroid diseases by the Chinese Medical Association, the prevalence of thyroid nodules is as high as 18.6%, and malignant nodules (thyroid cancer) account for 5%-15%. In order to obtain the correct treatment for patients with thyroid nodules, it is very important to accurately distinguish between benign and malignant nodules. There are two commonly used methods for diagnosing benign and malignant thyroid nodules: non-invasive thyroid ultrasound imaging and invasive fine-needle aspiration biopsy (fine-needle aspiration biopsy, FNAB). FNAB is the gold standard for nodule diagnosis, but mass screening using FNAB is traumatic and costly for patients. ...

AI Technical Summary

Problems solved by technology

Although these studies have made different attempts to extract local and global features, they are also limited to changing the input of the neural network to obtain features under different views.

Images