Datasets for your Research
Download our datasets
As part of MVTec's commitment for advancing machine vision research, MVTec offers several datasets to download. Please refer to the following for more details and references.
MVTec Industrial 3D Object Detection Dataset (MVTec ITODD)
Dataset for benchmarking 3D object detection methods focusing on industrial scenarios. It contains range images and grayscale images of several object classes that are frequently found in industrial setups.
Toolkit for Measuring the Accuracy of Object Trackers
The easy-to-use toolkit for evaluating the accuracy of trackers includes upper bounds for all box-base trackers on the Visual Object Tracking (VOT) and Densely Annotated Video Segmentation (DAVIS) challenges.
MVTec Densely Segmented Supermarket Dataset (MVTec D2S)
Dataset for benchmarking instance segmentation or object detection methods in the industrial setting of a groceries store. It contains objects from 60 different categories captured in various poses and on different backgrounds. Each scene is captured under three different lightings and ten rotations.
MVTec Anomaly Detection Dataset (MVTec AD)
Dataset for benchmarking anomaly detection algorithms. It contains images from 15 different object and texture categories. Each category consists of defect-free training images, as well as test images that contain various types of defects. Furthermore, pixel-precise ground truth annotations of the defects are provided.