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

CT intracranial hemorrhage detection system based on dynamic map loss neural network

A technology of dynamic map and intracranial hemorrhage, applied in the direction of biological neural network model, neural architecture, neural learning method, etc., can solve the problems of accurate statistics of bleeding volume, manual labeling deviation, and inability to accurately calculate bleeding volume, etc., to achieve accurate Bleeding volume statistics, the effect of reducing model deviation

Active Publication Date: 2021-04-13
SHAN DONG MSUN HEALTH TECH GRP CO LTD
View PDF10 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

In clinical practice, it is difficult for doctors to accurately count the amount of bleeding, and small bleeding points may be neglected by junior doctors
In the field of artificial intelligence, deep learning algorithms have been used to detect intracranial hemorrhage. However, most of the existing methods cannot accurately determine the labeling boundary pixel by pixel due to the deviation of manual labeling, and cannot accurately calculate the amount of bleeding.

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
  • CT intracranial hemorrhage detection system based on dynamic map loss neural network
  • CT intracranial hemorrhage detection system based on dynamic map loss neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0055] CT intracranial hemorrhage detection system based on dynamic map loss neural network, including:

[0056] The data set acquisition module is configured to acquire brain CT image data, and mark intracranial hemorrhage mask and background;

[0057] The feature extraction module is configured to perform convolution operations and maximum pooling operations in multiple cycles to obtain each feature output map;

[0058] The calculation feature extraction module is configured to perform a convolution operation on the feature output map or the joint feature and then perform a deconvolution operation to obtain a corresponding calculation feature map;

[0059] The joint module is configured to perform a stacking operation on the calculated feature map and the different times of the feature output map after the cutting operation to obtain the corresponding joint feature;

[0060] The segmentation module is configured to perform a convolution operation on the final feature output i...

Embodiment 2

[0096] The specific workflow includes the following steps:

[0097] a) Construct an intracranial hemorrhage segmentation dataset: collect brain CT image data, and label the intracranial hemorrhage mask and background.

[0098] b) Input brain CT image X into convolution module C 1 , using the computer through the convolutional layer C 1 Two 3-dimensional convolution operations are processed to obtain the feature output map C 1 (X);

[0099] c) Use the computer to output the features in Figure C 1 (X) Perform maximum pooling operation, compress feature map C 1 (X), get the updated feature output map C' 1 (X);

[0100] d) Output the updated feature map C' 1 (X) Input convolution module C 2 , using the computer through the convolutional layer C 2 Two 3-dimensional convolution operations are processed to obtain the feature output map C 2 (X);

[0101] e) Use the computer to output the features in Figure C 2 (X) Perform maximum pooling operation, compress feature map C ...

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 provides a CT intracranial hemorrhage detection system based on a dynamic map loss neural network, which uses a dynamic map loss function to reduce model deviation caused by wrong annotation at a mask boundary, and dynamically adjusts the weight of a pixel value annotated at the mask boundary during loss calculation in cooperation with a U-net network; therefore, the model can learn knowledge which should be learned, and ignores possible wrong knowledge at the boundary. According to the invention, the weight of the bleeding area edge participating in loss calculation can be dynamically adjusted, and the influence of edge error mark pixels on model adjustment is reduced. Therefore, the model can accurately fit a bleeding area and accurately calculate the bleeding amount.

Description

technical field [0001] The invention belongs to the technical field of intracranial hemorrhage detection, in particular to a CT intracranial hemorrhage detection system based on a dynamic map loss neural network. Background technique [0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art. [0003] Intracranial hemorrhage, which refers to bleeding that occurs inside the brain, is a serious health problem that requires prompt, sometimes often intensive, treatment. Determining the site and type of bleeding is a critical step in treating a patient. In clinical practice, it is difficult for doctors to accurately count the amount of bleeding, and small bleeding points may be neglected by junior doctors. In the field of artificial intelligence, deep learning algorithms have been used to detect intracranial hemorrhage. However, most existing methods cannot accurately determine the l...

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): G06T7/00G06T7/12G06K9/46G06N3/04G06N3/08
CPCG06T7/0012G06T7/12G06N3/08G06T2207/10081G06T2207/20081G06T2207/30016G06T2207/30104G06V10/44G06N3/045Y02T10/40
Inventor 樊昭磊吴军曲荣芳颜红建尚永生
Owner SHAN DONG MSUN HEALTH TECH GRP CO LTD
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