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

Intelligent image annotation method

A technology of intelligent images and image files, applied in image enhancement, image analysis, image data processing, etc., can solve the problems of low data labeling efficiency, achieve the effects of shortening labeling time, improving labeling work, and improving efficiency

Pending Publication Date: 2021-06-04
中科海拓(无锡)科技有限公司
View PDF8 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] The main purpose of the present invention is to provide an intelligent image labeling method, which can effectively solve the problem of low data labeling efficiency in the background technology

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
  • Intelligent image annotation method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0032] Such as figure 1 As shown, an intelligent image labeling method includes the following steps:

[0033] Step 1: Obtain a database of image files to be marked, and select an image file as an OK sample;

[0034] Step 2: judge whether the match with the OK sample is successful;

[0035] Step 3: Manually compare the similarity between the matched file and the OK sample file;

[0036] Step 4: If the threshold value is appropriate and no adjustment is required, continue the previous operation; if the threshold value is not appropriate, it needs to be adjusted before repeating the previous operation, and then operate according to the similarity.

[0037] Step 1 includes the following sub-steps:

[0038] The first step: scramble the unlabeled data set, and then perform random extraction to extract a certain number of image data files;

[0039] The second step: mark the relevant defects on the extracted image data files;

[0040] The third step: classify the marked image dat...

Embodiment 2

[0051] Such as figure 1 As shown, an intelligent image labeling method includes the following steps:

[0052] Step 1: Obtain a database of image files to be marked, and select an image file as an OK sample;

[0053] Step 2: judge whether the match with the OK sample is successful;

[0054] Step 3: Manually compare the similarity between the matched file and the OK sample file;

[0055] Step 4: If the threshold value is appropriate and no adjustment is required, continue the previous operation; if the threshold value is not appropriate, it needs to be adjusted before repeating the previous operation, and then operate according to the similarity.

[0056] Step 1 includes the following sub-steps:

[0057] Step 1: Scramble the unlabeled data set, and then perform random extraction to extract a certain number of image data files; for example, if a certain amount of image data is extracted, then the remaining data sets can be set as samples to be extracted pond;

[0058] The se...

Embodiment 3

[0080] Such as figure 1 As shown, an intelligent image labeling method includes the following steps:

[0081] Step 1: Obtain a database of image files to be marked, and select an image file as an OK sample;

[0082] Step 2: judge whether the match with the OK sample is successful;

[0083] Step 3: Manually compare the similarity between the matched file and the OK sample file;

[0084] Step 4: If the threshold value is appropriate and no adjustment is required, continue the previous operation; if the threshold value is not appropriate, it needs to be adjusted before repeating the previous operation, and then operate according to the similarity.

[0085] In step 2, judge whether the match with the OK sample is successful. The specific operation is that you only need to run the script tool_retrieve.py, and follow the samples in the sample folder in one step, and automatically cut to the target folder.

[0086] In step 2, match the image data to be labeled with the sample data...

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 intelligent image annotation method, which comprises the following steps of: 1, obtaining a to-be-annotated image file database, and selecting an image file as an OK sample; 2, judging whether the image file is successfully matched with an OK sample or not; 3, manually comparing the similarity between the matched file and the OK sample file; 4, if the threshold value is appropriate and does not need to be adjusted, continuing to execute the previous operation; if the threshold value is not appropriate, repeating the previous operation after adjustment, and then performing the operation according to the similarity. The intelligent image labeling method is suitable for the characteristic of target classification of the industrial image data, overcomes the problems of large data volume, long labeling time, insufficient training time and the like of the industrial image data in manual labeling, and can better perform labeling work of the industrial image data.

Description

technical field [0001] The invention relates to the field of artificial intelligence, in particular to an intelligent image labeling method. Background technique [0002] Image annotation plays a pivotal role in industrial vision. The main goal of image annotation is to complete the labeling of specific labels required by the task. This can include labeling of image segmentation, labeling of image recognition, and labeling of image defects; among them, the work of labeling image segmentation and image recognition is to use labeling software to draw labels on images (segmentation is generally marked with borders, recognition Usually irregular polygons), sometimes even labeling pixel-level labels. [0003] The general task requirements of computer vision are the following: object detection, line / edge detection, segmentation, key point recognition, image classification. Object detection generally uses two types of 2D detection and 3D detection. Among them, 2D detection is u...

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): G06K9/62G06T7/00
CPCG06T7/0004G06T2207/20081G06F18/22G06F18/214
Inventor 谢传鹏程坦刘涛吕剑
Owner 中科海拓(无锡)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

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com