From Scarcity to Abundance: Expansion Manufacturing Data through Limited Defect Images

Simple item page
dc.contributor.authorMoon, Junhyung
dc.contributor.authorYang, Minyeol
dc.contributor.authorPark, Songmi
dc.contributor.authorJeong, Jongpil
dc.date.accessioned2023-12-26T18:36:38Z
dc.date.available2023-12-26T18:36:38Z
dc.date.issued2024-01-03
dc.identifier.doihttps://doi.org/10.24251/HICSS.2024.124
dc.identifier.isbn978-0-9981331-7-1
dc.identifier.other0284f37e-fad8-4d96-9be2-22467979f0ac
dc.identifier.urihttps://hdl.handle.net/10125/106501
dc.language.isoeng
dc.relation.ispartofProceedings of the 57th Hawaii International Conference on System Sciences
dc.rightsAttribution-NonCommercial-NoDerivatives 4.0 International
dc.rights.urihttps://creativecommons.org/licenses/by-nc-nd/4.0/
dc.subjectData-driven Services and Servitization in Manufacturing: Innovation, Engineering, Transformation, and Management
dc.subjectdata augmentation
dc.subjectdata scarcity
dc.subjectgenerative ai
dc.subjectsmart manufacturing
dc.titleFrom Scarcity to Abundance: Expansion Manufacturing Data through Limited Defect Images
dc.typeConference Paper
dc.type.dcmiText
dcterms.abstractThe increasing adoption of IoT sensors, communication capabilities, and software applications in manufacturing environments has led to a growing demand for handling diverse large-scale manufacturing data. This trend indicates that AI is being researched and developed as an essential tool for improving cost-effectiveness and efficiency. Recently, there has been a significant increase in demand for process improvement using deep learning technology in smart manufacturing processes. However, obtaining a sufficient amount of training data in real industrial environments is challenging due to security and cost concerns for companies. Therefore, we propose utilizing generative artificial intelligence to efficiently expand manufacturing datasets. For data augmentation, we use a model that combines Stable Diffusion and LoRA fine tuning, and apply the text generation approach of BLIP. We anticipate that these data augmentation will help to improve the performance of artificial intelligence in the manufacturing field while reducing the cost of data collection.
dcterms.extent10 pages
prism.startingpage1027

Files

Original bundle

Now showing 1 - 1 of 1
Thumbnail Image
Name:
0100.pdf
Size:
2.78 MB
Format:
Adobe Portable Document Format
Download

Collections

Data-driven Services and Servitization in Manufacturing: Innovation, Engineering, Transformation, and Management