Eye image processing model construction method and device

An image processing and construction method technology, which is applied in the field of image processing to reduce interference, achieve positioning and visualization, and improve research and judgment capabilities.

Pending Publication Date: 2019-12-20
SHENZHEN UNIV
View PDF4 Cites 11 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Existing automatic ROP screening methods only give a diagnostic result, and cannot further provide physicians with more identification support

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
  • Eye image processing model construction method and device
  • Eye image processing model construction method and device
  • Eye image processing model construction method and device

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0024] This embodiment provides as figure 1 A method for building an eye image processing model is shown, including:

[0025] S1. Setting the residual network as the basic processing model;

[0026] S2. Add a feature detection module based on the attention mechanism at the end of the residual block of the residual network to obtain a classification model;

[0027] S3, training the classification model based on the ROP picture;

[0028] S4. Process the classification model based on the weighted gradient class activation mapping, realize the location and visualization of pathological parts, and output corresponding pathological images and / or type information.

[0029] The specific processing model construction principles include:

[0030] Using ResNet50 as the basic image processing neural network, a feature detection module including channel attention and spatial attention is used to enhance the feature representation ability of the neural network, so that the neural network...

Embodiment 2

[0055] This embodiment provides as figure 2 A kind of eye image processing model building device comprises:

[0056] The initial setting unit 1 is used to set the residual network as the basic processing model;

[0057] The modification unit 2 is used to add a feature detection module based on the attention mechanism at the end of the residual block of the residual network to obtain a classification model;

[0058] Training unit 3, for training classification model based on ROP pictures;

[0059] The visualization unit 4 is configured to process the classification model based on the weighted gradient class activation mapping, realize the location and visualization of pathological parts, and output corresponding pathological images and / or type information.

[0060] This embodiment provides an eye image processing model, including: a residual network, a feature detection module, and a weighted gradient-like activation mapping module; the feature detection module is connected ...

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 eye image processing model construction method and device. The method comprises the following steps: setting a residual network as a basic processing model; adding a featuredetection module at the tail end of the residual block to obtain a classification model; training the classification model based on ROP pictures; and processing the classification model based on weighted gradient class activation mapping, realizing positioning and visualization of a pathological part, and outputting corresponding pathological image and / or type information. The device is used forexecuting the method. By setting a basic processing model, a feature detection module is added at the tail end of the residual block. Interference of non-target features can be reduced through an attention mechanism, and the recognition efficiency is improved. Training a classification model based on the ROP pictures to define an applicable range; based on the weighted gradient class activation mapping processing classification model, positioning and visualization of pathological parts are achieved, corresponding pathological images and / or type information are / is output, the pathological structure can be clearly displayed, and the research and judgment capacity of doctors for specific symptoms can be improved.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to a method and device for constructing an eye image processing model. Background technique [0002] Retinopathy of prematurity (ROP) is a retinal vascular proliferative disease, mainly seen in premature and low birth weight infants. ROP is the leading cause of childhood blindness. Early screening and prompt treatment are critical to preventing ROP blindness. Due to factors such as heavy screening workload and insufficient professional ophthalmologists, research on automatic ROP screening methods is expected to reduce the burden on doctors and has certain clinical value. The existing automatic ROP screening method only gives a diagnosis result, and cannot further provide doctors with more identification support. Contents of the invention [0003] The embodiments of the present invention aim to solve one of the technical problems in the related art at least to a certain...

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): G06K9/00G06K9/62G06N3/08G06T7/00G16H30/00G16H50/20
CPCG06T7/0012G06N3/08G16H30/00G16H50/20G06T2207/30041G06V40/193G06V40/18G06F18/24G06F18/214
Inventor 雷柏英黄珊张国明汪建涛曾键赵金凤
Owner SHENZHEN 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