FrontierThe story, in brief
Novel “Kaputt” dataset sets new benchmark for large-scale visual defect detection
238,000 images. Amazon's new 'Kaputt' dataset just reset the benchmark for AI-powered defect detection in retail.
The KeyNews take
Why it matters
Amazon Science released a large-scale visual defect detection dataset that advances machine learning capabilities for real-world retail quality control. This benchmarking effort signals enterprise AI's shift toward solving tangible operational problems at scale.
The key facts
5 to know238,000+ images in 'Kaputt' dataset
Benchmark advancement in visual defect detection
Complex retail applications focus
Published by Amazon Science (October 2025)
Large-scale dataset for ML training and evaluation
Go to the source
Amazon Scienceamazon.science
Publisher excerpt: A new dataset with over 238,000 images challenges and advances the state of the art in visual defect detection for complex retail applications.