A benign and malignant classification system for ultrasonic thyroid nodules based on transfer learning and feature fusion

Abstract

A classification method for benign and malignant ultrasound thyroid nodules based on transfer learning and feature fusion. Firstly, the ultrasound image is preprocessed and scaled to a uniform size, then the traditional underlying features are extracted from the ultrasound image, and then the natural image The model trained by the deep neural network is used to extract the high-level semantic features of the ultrasound image. Afterwards, the low-level features are fused with the high-level features, and the final feature vector is obtained by feature screening using the degree of discrimination between benign and malignant thyroid nodules. Train the support vector machine classifier to classify the benign and malignant thyroid nodules; the present invention integrates the low-level features and high-level features, and performs significant feature screening, making up for the ability of a single feature to describe the characteristics of thyroid nodules at the semantic level The classification accuracy is effectively improved; by introducing transfer learning, it solves the problem that there are few medical sample images and cannot be directly trained to obtain deep features.

Description

technical field [0001] The invention belongs to the technical field of medical image classification, is suitable for ultrasonic thyroid classification, and specifically relates to an ultrasonic thyroid nodule benign and malignant classification system based on migration learning and feature fusion. Background technique [0002] Ultrasonography is one of the most valuable diagnostic methods for thyroid nodules. When ultrasonic waves propagate in the human body, due to the different acoustic impedance and attenuation characteristics of different tissues, they show different echo intensities. Benign and malignant nodules appear differently in ultrasound images, so image recognition and image classification methods can be used to automatically classify and distinguish benign and malignant thyroid nodules. This method can provide doctors with auxiliary diagnostic means, reduce the pressure of doctors' clinical diagnosis, and solve the problem of over-reliance on doctors' subject...

AI Technical Summary

Problems solved by technology

The difficulty of introducing this feature into ultrasound image classification is that in the medical field, such a large data set cannot be obtained to train a targeted deep network, and a small-scale medical image data set (usually around hundreds) is used to train a network containing With a deep network of millions of parameters, it is foreseeable that the training will fall into overfitting

Images