Convolutional neural network model for predicting and generating NBI image according to endoscope white light image and construction method and application of convolutional neural network model

A convolutional neural network and construction method technology, applied in biological neural network models, neural architectures, applications, etc., can solve problems such as inability to effectively obtain deep semantic information, affecting the accuracy of image analysis, etc.

Pending Publication Date: 2020-10-30
BEIJING CHAOYANG HOSPITAL CAPITAL MEDICAL UNIV
View PDF3 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, these methods based on manual feature extraction have great limitations and cannot effectively obtain more abstract deep-level semantic information, thus affecting the accuracy of image analysis

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
  • Convolutional neural network model for predicting and generating NBI image according to endoscope white light image and construction method and application of convolutional neural network model
  • Convolutional neural network model for predicting and generating NBI image according to endoscope white light image and construction method and application of convolutional neural network model
  • Convolutional neural network model for predicting and generating NBI image according to endoscope white light image and construction method and application of convolutional neural network model

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0046] The method for constructing the convolutional neural network model for predicting and generating NBI images of the present embodiment, the steps are as follows:

[0047] 1. Establish a neural network structure

[0048] This embodiment establishes as figure 1 The neural network structure shown, the network structure consists of four parts: input layer, pre-trained encoder, decoder, output layer.

[0049] In the CNN model structure, the shallow network can extract geometric information such as corners and shapes of the image, and the deep network can extract high-order semantic information; in order to make full use of this feature of CNN and solve the problem of "small samples" of medical images The problem is to use deconvolution and upsampling techniques to let the feature matrix of each layer of the encoder and the corresponding layer of the decoder perform feature fusion (feature matrix addition), so that the decoder not only retains low-order geometric information ...

Embodiment 2

[0062] Such as image 3 As shown, the construction method of the convolutional neural network model for predicting and generating NBI images of the present embodiment is improved as follows on the basis of the construction method of Embodiment 1:

[0063] In step 1, in establishing the neural network structure, an additional head network for benign and malignant classification is added to the underlying shared feature map, and the final output layer is divided into two branches, which predict and generate NBI images and classify benign and malignant, respectively, so that Give the model more capabilities.

[0064] In step 3 of model training, the initial model is also trained using data sets including multiple pathological results, and a convolutional neural network model for predicting and generating NBI images and classifying benign and malignant lesion images is constructed.

[0065] Step 4 model application, such as Figure 4 As shown, the white light image to be analyze...

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 convolutional neural network model for predicting and generating an NBI image according to an endoscope white light image and a construction method and application of the convolutional neural network model. The construction method comprises the following steps: S1, establishing a neural network structure comprising an input layer, an encoder, a decoder and an output layer;S2, performing assignment on the neural network structure by adopting a fusion transfer learning algorithm to obtain an initial model; S3, training the initial model by adopting a data set comprisinga plurality of pairs of paired white light images and NBI images. The convolutional neural network model can predict and generate the NBI image and analyze the lesion property according to the endoscope white light image without the help of NBI equipment, the real-time prediction speed requirement is met on the premise that the accuracy is guaranteed, and a new thought is provided for image analysis of a digestive endoscope in clinic.

Description

technical field [0001] The present invention relates to the technical field of convolutional neural networks, in particular to a convolutional neural network model for predicting and generating NBI images based on endoscopic white light images and its construction method and application. Background technique [0002] Gastrointestinal tumors have a high incidence and great harm, so early diagnosis and treatment are particularly important. However, the early manifestations of gastrointestinal tumor lesions are varied, and it is difficult to identify and diagnose with traditional white light endoscopy, and sometimes it is difficult to determine the boundary and depth of invasion. [0003] Studies have shown that compared with white light imaging, endoscopic narrow band imaging (NBI) combined with magnifying endoscopy has higher sensitivity and specificity for the diagnosis of early gastric cancer. At present, endoscopists mostly use NBI to observe gastrointestinal lesions. NB...

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): G06T7/00G06K9/62G16H30/40G06N3/04A61B1/00
CPCG06T7/0012G16H30/40A61B1/00032G06T2207/30096G06N3/045G06F18/214G06F18/253
Inventor 王泽楠刘明刘心娟郝建宇
Owner BEIJING CHAOYANG HOSPITAL CAPITAL MEDICAL UNIV
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