Data annotation, model training and image processing method and device and storage medium

An image and model technology, applied in the field of image processing, can solve the problems of missing labeling, difficulty in finding with naked eyes, and difficulty in labeling work.

Active Publication Date: 2020-04-17
北京医准智能科技有限公司
View PDF7 Cites 20 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

But one of the more difficult types is diffuse multiple lesions, which often occupy most of the lung space, and the thousands of lesions make labeling difficult
There are three reasons for this. 1. Due to the large number of lesions, finding and labeling all lesions will bring the risk of missing labels (such as diffuse infection of the whole lung, such as diffuse increase...

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
  • Data annotation, model training and image processing method and device and storage medium
  • Data annotation, model training and image processing method and device and storage medium
  • Data annotation, model training and image processing method and device and storage medium

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0056] figure 2 A schematic block diagram of a data labeling method according to an embodiment of the present invention is shown.

[0057] Such as figure 2 As shown, according to the data labeling method of an embodiment of the present invention, this embodiment takes object detection in medical images as an example to illustrate the inventive solution, and the method of the present invention is not limited to the processing of medical images. In addition, in the embodiment of the present invention, the Faster R-CNN model is taken as an example to illustrate the neural network model used for target detection. The neural network model in the present invention is not limited to this, and various neural networks with different structures can be used in the present invention The model and the data labeled using the standard method of the present invention are trained to improve its detection effect on specific types of images and targets.

[0058] A sample data labeling method...

Embodiment 2

[0083] image 3 A schematic block diagram according to another embodiment of the present invention is shown.

[0084] Such as image 3 As shown, the second embodiment of the present invention provides a kind of training method for the neural network of object detection, and the neural network in the present embodiment generally comprises four parts: feature extraction network (convolutional layer), area generation network ( Region Proposal Network, RPN), target region pooling network (ROI pooling layer) and target classification network (Classification), typically, use the frame of Faster R-CNN model to construct the neural network of target detection in the present embodiment, Faster R -CNN model is an existing technology, a typical Faster R-CNN model is attached figure 1 As shown, the structure of an exemplary neural network in the present invention is briefly introduced below, and will not be elaborated again:

[0085] Among them, the feature extraction network or featur...

Embodiment 3

[0119] as attached Figure 5 As shown, Embodiment 3 of the present invention provides an image processing method using a neural network model to detect objects in an image. The method specifically includes the following steps:

[0120] S401: Acquire an image to be processed.

[0121] The image to be processed is an image of the same type as the sample image used for neural network training, and the target to be detected is the same as the target category marked in the sample image.

[0122] Same as the previous embodiment, this embodiment still takes the human lung image as an example, and the detection target is a lung nodule. The protection scope of the present invention is not limited thereto, and the data labeling method, neural network training method and image processing method of the present invention can be used to detect various types of images and targets according to different labeling data.

[0123] S402: Invoke a neural network model to perform target detection...

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 lung focus detection algorithm based on weak annotation data. The algorithm comprises the steps of data labeling, model training and image processing, wherein a partially labeled weak data labeling method is adopted for labeling, a labeling result is perfected by referring to the thought of difficult case mining in the data labeling process, and it is guaranteed that a high-quality labeled data set is provided for subsequent model training under the condition that only part of focuses are labeled. According to the method, model training is combined with a 'weak data 'marking mode, positive and negative sample anchor points for network training are acquired from positive and negative sample sets respectively, and the focal loss hyper-parameters are adjusted according to the source characteristics of the positive and negative sample anchor points, so that the detection model obtains a better detection effect.

Description

technical field [0001] The invention relates to the field of image processing, in particular to a training method for an image processing model and a method for processing images using the model. Background technique [0002] Lung cancer is one of the malignant tumors with the fastest growing morbidity and mortality, which has posed a great threat to the health and life of the population. The survival rate of lung cancer is highly correlated with the stage of the disease when it is first diagnosed. If it can be detected at an early stage, the 5-year survival rate can reach 70% to 90%. Compared with other cancers, the biological characteristics of lung cancer are very complex, and most of them have no obvious symptoms in the early stage. Nearly 75% of the people are found in the middle and late stages, and the treatment cost is high and the effect is not good. Therefore, early detection and diagnosis of lung cancer is particularly important. Lung cancer usually presents as ...

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): G06T7/00
CPCG06T7/0012G06T2207/10081G06T2207/20081G06T2207/20084G06T2207/30061G06T2207/30096G06T2207/30204
Inventor 张佳琦王子腾孙安澜吕晨翀丁佳胡阳
Owner 北京医准智能科技有限公司
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