The MVTec 3D-AD Dataset for Unsupervised 3D Anomaly Detection and Localization
This dataset addresses the problem of detecting defects in manufactured products for industrial inspection, but it is incremental as it extends existing 2D anomaly detection datasets to 3D.
The authors tackled the lack of a comprehensive 3D dataset for unsupervised anomaly detection and localization by introducing the MVTec 3D-AD dataset, which includes 10 object categories with anomaly-free training data and defective test samples, and initial benchmarks show significant room for improvement.
We introduce the first comprehensive 3D dataset for the task of unsupervised anomaly detection and localization. It is inspired by real-world visual inspection scenarios in which a model has to detect various types of defects on manufactured products, even if it is trained only on anomaly-free data. There are defects that manifest themselves as anomalies in the geometric structure of an object. These cause significant deviations in a 3D representation of the data. We employed a high-resolution industrial 3D sensor to acquire depth scans of 10 different object categories. For all object categories, we present a training and validation set, each of which solely consists of scans of anomaly-free samples. The corresponding test sets contain samples showing various defects such as scratches, dents, holes, contaminations, or deformations. Precise ground-truth annotations are provided for every anomalous test sample. An initial benchmark of 3D anomaly detection methods on our dataset indicates a considerable room for improvement.