Simultaneous object detection and tracking using center points.
MIT License
2.38k
stars
526
forks
source link
What kind of benchmarks are needed? What about UA-DETRAC? "The models are not evaluated on any benchmarks since there are no suitable ones in this setting. " #61
oth models are trained with the "training on static image data" technic in our paper.
The models are not evaluated on any benchmarks since there are no suitable ones in this setting. We provide them for demo purpose only.
请问这是因为COCO是悬在半空的卡口相机拍的画面,所以车载相机拍的跟踪数据集不合适吗?UA-DETRAC是卡口相机拍的。
Many thanks!