- Media recognition
Unsupervised pre-training method for industrial images
We built a dedicated model for industrial fields using small number of real images, for highly accurate analysis even with specialized images.
- Images captured under special conditions or with specialized equipment often involve significant time and cost, resulting in a smaller scale of available images.
- In some cases, pre-training involves using a large natural dataset consisting of images captured with conventional cameras cannot be effectively used for industrial images.
- We built a pre-training model for industrial images by deliberately generating artificial images that include local structures within target images.
Applications
- Visual inspections of products on manufacturing lines
- Inspection of biological images
Benchmarks, strengths, and track record
- We evaluated the accuracy of this AI using five publicly available non-natural image datasets (infrared, microscopic, wafer, pathological, and fundus images).
For each dataset, a small number of images, ranging from 40 to 1,000, were randomly selected to generate between 9,000 and 30,000 pre-training images for the image classification task.
As a result of the evaluation, using this technology for pre-training achieved higher accuracy than pre-training using ImageNet, a typical large-scale natural image dataset that contains 1.3 million images.
Inquiries
Inquiries to Toshiba Corporate Laboratory (Komukai region)
Please include the title “Toshiba AI Technology Catalog: Unsupervised pre-training method for industrial images” or the URL in the inquiry text.
Please note that because this technology is currently the subject of R&D activities, immediate responses to inquiries may not be possible.
