Datasets

ADAPT JR-Sim2Real Dataset
This synthetic training dataset was created specifically for the ADAPT sim2real Object Detection Challenge 2023. In order to create a comprehensive dataset, we were provided with 12 CAD models which allowed us to create a variety of data samples under different lighting and other environmental conditions.
Learning Scene Semantics from Vehicle-centric Data for City-scale Digital Twins
Dataset for "Learning Scene Semantics from Vehicle-centric Data for City-scale Digital Twins", Fürntratt et al. Data are anonymized and provided with segmentation mask ground truths.
AnonCityDT: Anonymisation Dataset for Cityscape Digial Twins
The datasets folder contains lists of files of images from ADE20K, Cityscapes and Places365, with annotations for segmenting human heads and license plates. The tunnel_transform folder contains the code used for data augmentation in order to improve the performance on images obtained from undistorting fisheye camera images.