CVPR2023

ARKitTrack: A New Diverse Dataset for Tracking Using Mobile RGB-D Data

Haojie Zhao, Junsong Chen, Lijun Wang, Huchuan Lu

Abstract

Figure 1 . Samples from ARKitTrack. We capture both indoor and outdoor sequences (1st row) in many scenes, including zoo, market, office, square, corridor, etc. Lots of scenarios are presented in our dataset, e.g., low or high light conditions (2nd row), surrounding clutter (3rd row), out-of-plane rotation, motion blur, deformation, etc. (4th row). Besides, we annotate each frame with object masks.