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

Computer-aided system and method for detecting mammary molybdenum target lumps through data driving

A computer-aided, data-driven technology, applied in computer parts, computing, instruments, etc., can solve the problems of lack of medical principle support for training models, low model recognition rate, large data requirements, etc., to improve generalization ability and recognition rate. , the effect of reducing training data and simple model structure

Pending Publication Date: 2019-05-24
青岛中科智康医疗科技有限公司
View PDF4 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] The purpose of the present invention is: the present invention provides a computer-aided system and method for data-driven mammary gland mass detection, which solves the problem of high cost due to the large data demand when training the model in the existing method, and the lack of effective medical principle support for the training model. Low recognition rate and poor generalization ability

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
  • Computer-aided system and method for detecting mammary molybdenum target lumps through data driving
  • Computer-aided system and method for detecting mammary molybdenum target lumps through data driving
  • Computer-aided system and method for detecting mammary molybdenum target lumps through data driving

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0135] A data-driven computer-aided system for mammography mass detection comprising

[0136] The image processor is used to label and classify the collected clinical images, and the images are divided into a training set, a verification set and a test set;

[0137] ROI extractor for extracting the mass region in the processed image;

[0138] Data-driven model, used to build and train extractors that are good at extracting tumor area shape, edge texture, and density features;

[0139] The comprehensive recognition model is used to construct the CNN network to extract the overall features of the tumor, and then fuse the features obtained by the extractor corresponding to the data-driven model with the overall features of the tumor, and use the training set and verification set for training and verification to obtain the best Comprehensive recognition model, input the test set data into the best comprehensive recognition model for recognition, and obtain the detection results. ...

Embodiment 2

[0149] Based on Example 1, the data-driven model includes a shape feature extractor, an edge texture feature extractor, and a density feature extractor, the details of which are as follows:

[0150] The shape feature extractor includes an edge extraction unit for extracting the edge of the tumor area, an edge connection unit for connecting the extracted edges to obtain a closed edge, a filling unit for filling the closed edge with a solid color to obtain the shape of the tumor, and a unit for using the above-mentioned unit The processed training set and verification set are used to train and verify the constructed CNN network, and obtain a shape feature CNN model that can obtain the best shape features and is good at distinguishing three types of shapes: circular, oval, and irregular;

[0151] The shape feature extractor training includes the following steps:

[0152] Step a1: extracting edges in the mass region;

[0153] Step b1: connect the extracted edges to obtain closed ...

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 computer-aided system and method for data-driven breast molybdenum target lump detection, and relates to the field of auxiliary devices for lump detection. The system comprises an image processor, ROI extractor, Construction, The training comprises the following steps: extracting the shape of a lump region; Edge texture, a data driving model of the density feature extractor is used for constructing and extracting the overall features of the lumps; fusing characteristics obtained by an extractor corresponding to the data driving model, training and verifying by utilizing the training set and the verification set to obtain an optimal identification model, and inputting test set data into the optimal identification model for identification, thereby obtaining a comprehensive identification model of a detection result; According to the method, the problems of high cost caused by high training data requirement and low recognition accuracy caused by lack of a clinical medical principle in a training model in the prior art are solved, the recognition accuracy is improved through multi-feature fusion, the generalization ability and interpretability of the model areimproved through multi-feature driving recognition, the training data are reduced, and the cost is reduced.

Description

technical field [0001] The invention relates to the field of auxiliary devices for mass detection, in particular to a computer-aided system and method for data-driven mammography mass detection. Background technique [0002] At present, commonly used methods in the field of image recognition include methods based on traditional machine learning and methods based on deep learning; traditional machine learning methods manually construct features and complete image classification / detection tasks with the help of trained classifiers; deep learning-based The method is goal-oriented and uses deep learning to achieve end-to-end training of features and classifiers. The effect is remarkable and it is extremely suitable for the classification / detection of ordinary images. [0003] Mammography images are images obtained by taking X-rays of breasts. They are unnatural images, which are more complex than natural images. The mammography image recognition method based on deep learning has...

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): G06K9/62G06K9/34
Inventor 韩云翟红波
Owner 青岛中科智康医疗科技有限公司
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