Thyroid tumor ultrasonic image recognition method and device

A technology for thyroid tumors and ultrasound images, applied in the field of image recognition, can solve problems such as low recognition accuracy

Inactive Publication Date: 2018-09-11
FUDAN UNIV SHANGHAI CANCER CENT +1
View PDF3 Cites 27 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] In view of the above-mentioned problems in the prior art, the purpose of the present invention is to provide a method and device for ultrasonic image recognition of thyroid tumors that can effectively select tumor image regions to improve the recognition accuracy, and a method that can avoid cases due to image comparisons. Ultrasonic image recognition method and device for thyroid tumors with low recognition accuracy caused by few

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Thyroid tumor ultrasonic image recognition method and device
  • Thyroid tumor ultrasonic image recognition method and device
  • Thyroid tumor ultrasonic image recognition method and device

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0073] Example 1 Obtaining Ultrasonic Images of Thyroid Tumors

[0074] The original dataset was collected from patients with thyroid nodules by the Cancer Hospital of Fudan University from January 2014 to December 2016. All papillary thyroid carcinoma (PTC) lesions were pathologically confirmed by surgical resection specimens. Each image in the raw data set was ranked by three experienced radiologists using the Thyroid Imaging Reporting and Data System (TI-RADS). TI-RADS scores are divided into 2, 3, 4a, 4b, and 4c, respectively: no suspicion, probably benign nodules, a single suspicious feature, two suspicious features, and three or more suspicious features. Three radiologists also acquired sonographic features, including their composition, echogenicity, calcifications, margins, and shape.

[0075] The entire dataset has a total of 2836 original images, of which 1484 are PTC-diagnosed ultrasound images and 1352 are collected from benign tumors, including nodular diseases a...

Embodiment 2

[0076] Example 2 Tumor Region Selection, Margin Range Expansion

[0077] All images in the original data set are collected from the phenomenon reporting system. If the entire image is used as a training sample, its inherent background, text, etc. will greatly affect the extraction of benign and malignant features of tumors. Therefore, it is necessary to select the tumor area on the original image, that is, to use a rectangular frame to completely select the tumor block, such as figure 2 shown in the inner rectangle box in . The selection process is completed one by one by doctors using semi-automated selection scripts.

[0078] figure 2 The area shown in the middle and inner rectangular frame only includes the nodule itself, and the information of the surrounding tissue of the nodule is very necessary for the judgment of benign and malignant tumors, so the concept of margin range is added when cutting the image. figure 2 The outer rectangular frame in the two images is t...

Embodiment 3

[0082] Embodiment 3 image cutting, labeling form training set

[0083] The original data set was split according to the ratio of 6:1. The original data set has a total of 2836 images. In the experiment, 1275 PTC images and 1162 benign images were randomly selected as the training set, and the remaining 209 PTC images and 190 benign images were used as the test set. In the process of data set segmentation, it is noted that the images of the same patient can only exist in the training set or the test set at the same time, which prevents the interference of similar images.

[0084] In the process of experimental testing, the size of the tumor itself is also considered to be quite different, so the data set is divided into smaller than 0.5cm, 0.5cm to 1cm, and larger than 1cm. The distribution on Figure 4 As shown, the same distribution of training and testing data is satisfied.

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention discloses a thyroid tumor ultrasonic image recognition method and device, and the method comprises the steps: selecting a tumor region in a thyroid tumor ultrasonic image, extending a certain boundary scope and then performing cutting, carrying out the benign and malignant marking, and enabling the cut images to form a training set; training a selected CNN (Convolutional Neural Network) through the training set, and forming a thyroid tumor ultrasonic image recognition model; obtaining a to-be-recognized thyroid tumor ultrasonic image, selecting a tumor region and extending a certain boundary scope, and carrying out the benign and malignant recognition through the thyroid tumor ultrasonic image recognition model. The method and device are used for assisting a doctor to diagnose the benign and malignant thyroid tumors, obtain the accuracy of 90% or greater in the detection test of the benign and malignant thyroid tumors through the thyroid tumor ultrasonic image, and is ofgreat reference significance to the actual clinical diagnosis.

Description

technical field [0001] The present invention relates to the field of image recognition, in particular to a deep learning-based ultrasonic image recognition method and device for thyroid tumors. Background technique [0002] Epidemiological studies have shown that thyroid cancer is one of the most common malignant tumors in women, and its incidence has risen rapidly in recent years. Papillary thyroid carcinoma (PTC) is the main pathological type of thyroid cancer, accounting for a large proportion of the incidence. Ultrasound images are widely regarded as the main diagnostic tool for thyroid nodule screening and the preoperative evaluation method for PTC. Ultrasonographic features such as microcalcifications, high solid content, and irregular edges and shapes are typical for PTC considerations. The biggest limitation in the actual operation process is the dependence of the operator, and the accuracy of judgment of radiologists with different experience levels is different. ...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): G06T7/00A61B8/08
CPCA61B8/085A61B8/5215A61B8/5223G06T7/0012G06T2207/10132G06T2207/20081G06T2207/30096
Inventor 向俊卢宏涛官青王芬王蕴珺李端树杜佳俊秦宇
Owner FUDAN UNIV SHANGHAI CANCER CENT
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap