Weak label data denoising method based on regularized label propagation
A label propagation and data noise reduction technology, applied in the direction of instruments, character and pattern recognition, computer components, etc., can solve the problems of low accuracy and low efficiency of noise reduction, improve the efficiency of reduction, optimize the quality of labeling, and improve noise reduction The effect of accuracy
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment Construction
[0041] In the following, an example is used to describe the denoising method using the weakly labeled data denoising method based on regularized label propagation. as attached figure 1 as shown,
[0042] Step 1: Obtain the required weakly labeled sample dataset from the crowdsourcing data platform;
[0043] For large-scale data labeling, the crowdsourcing platform represented by Amazon (Amazon MechanicalTurk) is often used to distribute data on the Internet, labeling by network employees, and then combining multiple labeling results to give the final labeling of large-scale data sets . Although this labeling method makes good use of group wisdom and idle Internet resources, different employees have different professional fields and lack of expert experience, resulting in a certain proportion of data being incorrectly labeled. Large-scale weakly labeled data samples can be obtained through this crowdsourcing method that does not rely on domain expert experience. Although we...
PUM
Abstract
Description
Claims
Application Information
- R&D Engineer
- R&D Manager
- IP Professional
- Industry Leading Data Capabilities
- Powerful AI technology
- Patent DNA Extraction
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