Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

A lung texture recognition method based on deep neural network to extract apparent and geometric features

A geometric feature and appearance technology, applied in the field of lung texture recognition based on deep neural network to extract appearance and geometric features, can solve the problems of inability to complete high-precision texture recognition and ignore geometric features, so as to improve the accuracy of recognition , easy to build and enhance the effect of ability

Inactive Publication Date: 2020-06-16
DALIAN UNIV OF TECH
View PDF4 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

The above-mentioned deep learning-based methods only mine the apparent information of the lung texture, ignoring its geometric characteristics, thus failing to fully learn effective features, and unable to complete high-precision texture recognition

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
  • A lung texture recognition method based on deep neural network to extract apparent and geometric features
  • A lung texture recognition method based on deep neural network to extract apparent and geometric features
  • A lung texture recognition method based on deep neural network to extract apparent and geometric features

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0037] The present invention proposes a lung texture recognition method based on a deep neural network to extract apparent and geometric features, which is described in detail in conjunction with the accompanying drawings and embodiments as follows:

[0038] The present invention builds a dual-channel residual network, uses lung CT images for training, and achieves a high correct recognition rate in the test. The specific implementation process is as follows: figure 1 As shown, the method comprises the following steps;

[0039] 1) Prepare initial data:

[0040] 1-1) A total of 217 lung CT images of patients were collected in the experiment. Among them, the CT images of 187 patients contained 6 typical textures of diffuse lung disease, namely, nodular, emphysema, honeycomb, fixed, ground glass and ground glass with lines; the CT images of the remaining 30 patients In the image, only normal lung tissue texture is presented. The 217 CT images were used to generate 7 lung textu...

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 lung texture recognition method for extracting apparent and geometric features based on a deep neural network, which belongs to the fields of medical image processing and computer vision. Using 217 three-dimensional computed tomography (CT) images of the lungs as source data, several sets of data sets were obtained after preprocessing, where each set of data contained a small block of CT images, corresponding small blocks of geometric information images, and a category label. A dual-channel residual network framework is constructed, with CT image small blocks and corresponding geometric information small blocks as the input of each channel, and the apparent information and geometric information of the lung texture are learned through the dual-channel residual network, and it is effectively Fusion, so as to get a higher recognition rate. Furthermore, the proposed network has a clear structure, is easy to construct, and is easy to implement.

Description

technical field [0001] The invention belongs to the field of medical image processing and computer vision, and relates to using a deep neural network framework to extract the relevant features of the appearance and geometric information of lung computed tomography (CT) images, and to classify different types of lung CT image textures , specifically relates to a lung texture recognition method based on deep neural network to extract appearance and geometric features. Background technique [0002] Diffuse lung disease is a general term for lung diseases that present widely distributed large areas of lung shadows in the lung area of ​​CT images. Because these lung shadow textures are complex and easily confused, it is difficult to accurately identify different lung textures even for experienced radiologists. Therefore, it is necessary to establish a computer-aided diagnosis (CAD) system for accurate and efficient automatic recognition of lung texture in CT images of diffuse lu...

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 Patents(China)
IPC IPC(8): G06T7/00G06T7/40G06N3/04
CPCG06T7/0012G06T7/40G06T2207/30061G06T2207/20081G06T2207/10081G06N3/045G06T2207/20084G06T7/41G06N3/08G06N3/04
Inventor 徐睿叶昕辰丛臻
Owner DALIAN UNIV OF TECH
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com