Algorithm improvement of biomimetic pattern recognition in imaging pneumonia discrimination

A technology of bionic pattern recognition and pneumonia, applied in the field of image recognition, can solve the problems of decreased accuracy

Inactive Publication Date: 2019-08-02
电子科技大学成都学院
View PDF9 Cites 6 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] The discrimination of pneumonia type is of great significance to its treatment. Traditionally, computer-aided discriminat

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
  • Algorithm improvement of biomimetic pattern recognition in imaging pneumonia discrimination
  • Algorithm improvement of biomimetic pattern recognition in imaging pneumonia discrimination
  • Algorithm improvement of biomimetic pattern recognition in imaging pneumonia discrimination

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0101] according to image 3 The network structure is trained, and each image of various samples is input into the network for training and feature extraction, and finally a network model is formed. The specific parameters of the network are shown in Table 1. Among them, R represents the ReLU (linear rectification function) function, M represents the maximum pooling operation, and D represents Dropout.

[0102] Table 1 Neural Network Parameters

[0103]

[0104] The original image is convolved through this network. In the process of several convolutions, the image features are preserved and the invalid noise is reduced. Compared with the traditional feature extraction, the preserved features are more objective and effective.

Embodiment 2

[0106] The improved bionic pattern recognition constructed by the present invention and the traditional pattern recognition (selected SVM in this paper) all use the same training set, verification set and test set. There are three groups in the experiment, and the number of images included in each group is all different. A The group contains 1200 images, group B contains 2700 images, and group C contains 5850 images, but the ratio of training set, verification set and test set in each group is 8:1:1. The experiment verifies the effectiveness of the proposed scheme from the correct rate of judgment of various X-ray images, and finally analyzes the cause of the error recognition.

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 an algorithm improvement of biomimetic pattern recognition in imaging pneumonia discrimination. A healthy X-ray film, a viral pneumonia X-ray film and a bacterial pneumonia X-ray film are normalized into experimental images with the same size; the experimental image I after the normalization of the healthy X-ray film image is input into a convolutional neural network to extract an image feature vector; wherein the convolutional neural network comprises a convolutional layer, a pooling layer and a full connection layer. The image feature vector extraction method comprises the steps of A, carrying out convolution on an experimental image I and a convolution kernel, and outputting a result through an activation function to form a feature map 01 of the convolutional layer; carrying out convolution on the feature map 01 and a next convolution kernel, and outputting a result through an activation function to form a feature map 02 of a convolution layer; and the feature map 02 enters a pooling layer as an input image.

Description

technical field [0001] The invention belongs to the field of image recognition, and relates to the algorithm improvement of bionic pattern recognition in imaging pneumonia discrimination. Background technique [0002] Artificial intelligence has achieved many innovations in many industries, improving not only work efficiency, but also accuracy. As far as the medical field is concerned, artificial intelligence has been applied to clinical diagnosis problems, but the application of artificial intelligence to the field of medical imaging started relatively late. The reason is that the artificial intelligence used for medical imaging diagnosis must rely on human subjectivity to establish mathematical models. Support Vector Machine Classifiers Reduce Intra-Scan Variation in HRCT Classification of Regional Disease Patterns in Diffuse Lung Disease: Comparison with Bayesian Classifiers, A Support Vector Machine with Bayesian Based Algorithm Using 22 Quantitative Features To judge p...

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
IPC IPC(8): G06T7/00G06N3/04
CPCG06T7/0012G06T2207/30061G06T2207/10081G06T2207/20084G06T2207/20081G06N3/045
Inventor 邹倩颖吴宝永王小芳
Owner 电子科技大学成都学院
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