- Media recognition
- Anomaly detection
Technology for detecting anomalies and defects in images
Detects anomaly locations in images using only simple image-level annotations.
- This AI technology localizes anomalies in images using simple annotations with the label “anomaly/no anomaly” for each image.
- Our proposed method leverages residual features between input and reference (normal) images, enabling more accurate anomaly localization than conventional methods.
- The proposed method is applicable to various scenarios, including visual inspection of products, equipment, and infrastructure.
Applications
- All types of anomaly detection using images (various defects, such as cracks or rust)
Benchmarks, strengths, and track record
- Evaluations on three public datasets show that our method achieves higher localization accuracy for anomalies compared to existing weakly supervised approaches.
Inquiries
Inquiries to Toshiba Corporate Laboratory (Komukai region)
Please include the title “Toshiba AI Technology Catalog: Technology for detecting anomalies and defects in 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.
