NUS-PRO is the abbreviation for 'NUS People and Rigid Objects Dataset', which contains 365 image sequences with manually labeled bounding boxes and occlusion status. It is designed for evaluating algorithms of object tracking. To meet the demand for developing deep model, all annotations are released!  click the image for a better resolution Download Reference @article{nus_pro, title={{NUS-PRO: A New Visual Tracking Challenge}}, author={Li, A and Lin, M and Wu, Y and Yang, MH and Yan, S}, journal={ IEEE Transactions on Pattern Analysis and Machine Intelligence}, volume={38}, number={2}, pages={335--349}, year={2016} } |