The ParallelEye Dataset

We propose a new virtual image dataset called “ParallelEye”. For that we present a dataset generation pipeline that uses street map, computer graphics, virtual reality, and rule modeling technologies to construct a realistic, large-scale virtual urban traffic scene. The artificial scene matches the real world well in terms of fidelity and geographic information. In the artificial scene, we flexibly configure the camera (including its position, height, and orientation) and the environmental conditions, to collect diversified images. Each image has been annotated automatically with ground truth including semantic/instance segmentation, object bounding box, object tracking, optical flow, and depth.