Synthetic MVTec AD: Labeled Industrial Defect Images for Supervised Detection

AnywayLabs.ai created a fully synthetic dataset of defect images for industrial inspection, inspired by the MVTec AD dataset. The dataset reformulates anomaly detection as a supervised object detection task with labeled defect instances. It was last updated on 2026-04-27.

Use Cases

• Training supervised object detection models based on labeled defect instances.
• Benchmarking industrial inspection algorithms based on synthetic defect images.
• Developing and testing defect localization models based on the supervised task formulation.

Strengths

• Dataset is designed for a supervised object detection task, providing explicit labels for defect instances.
• Images are described as fully synthetic and focus on realistic industrial defects.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• The dataset is synthetic, which may not fully capture the complexity of real-world defect data.

Provenance

Source

AnywayLabs.ai

Collection Method

Synthetically generated, inspired by the original MVTec AD dataset.

Time Range

null

Freshness

Last updated 2026-04-27 12:44:36; freshness should be verified.

Geography

null