mmtracking
mmtracking copied to clipboard
OpenMMLab Video Perception Toolbox. It supports Video Object Detection (VID), Multiple Object Tracking (MOT), Single Object Tracking (SOT), Video Instance Segmentation (VIS) with a unified framework.
It would be useful if the repository included an example Colab notebook on how to train instance segmentation with your own dataset. The documentation is quite sparse about VIS, and...
1. Can we find out the inference output accuracy? 2. And if yes, so how?
RT,Whether mmtracking frame will support MOTR V2?
直接运行test.py deepsort_faster-rcnn_fpn_4e_mot17-public-half.py 得到结果没有问题,但是当我 运行train.py deepsort_faster-rcnn_fpn_4e_mot17-public-half.py 后得到文件 1-deepsort_faster-rcnn_fpn_4e_mot17-public-half.py 然后运行test.py 1-deepsort_faster-rcnn_fpn_4e_mot17-public-half.py --eval track get some error:KeyError: "'track_bboxes' not found in the outputs."怎么办
Hello! Last commit in this repo was in April 23. Is mmtracking currently maintained?
Can the existing project implement mask function for multi-Object tracking (instead of simple box selection of target objects).If not, how can I modify the existing code
Whether mmtracking and mmcv mmdection can improve compatibility between versions, mmtracking needs are too old
Hi, thanks for your great work. I've been using mmtracking for a while, and I'm pretty sure there is a bug in class SortTracker(). It's bug about implementation of retrieving...