The KITTI Vision Benchmark Suite

We present a simple yet effective quasi-dense
matching method to learn instance similarity from
hundreds of region proposals in a pair
of images. In the resulting feature space, a simple
nearest neighbor search can distinguish different
instances without bells and whistles. We apply our
method to joint object detection and tracking, along
with Faster R-CNN + ResNet50 without using location
or motion heuristics.

See the code for details.

@inproceedings{pang2021quasidense,
title={Quasi-Dense Similarity Learning for Multiple Object Tracking},
author={Pang, Jiangmiao and Qiu, Linlu and Li, Xia and Chen,
Haofeng and Li, Qi and Darrell, Trevor and Yu,
Fisher},
journal = {CVPR},
year = {2021}
}