.
675 |
--------------------------------------------------------------------------------
/README.md:
--------------------------------------------------------------------------------
1 | ## This is a placeholder for DAUMOT code release. The code will be released in the following days.
2 | ## Official Implementation for DAUMOT: Domain Adaptation for Unsupervised Multiple Object Tracking
3 | ** DAUMOT: Domain Adaptation for Unsupervised Multiple Object Tracking **
4 |
5 | The idea was originated from Guillaume Delorme and he implemented the code for FairMOT-DAUMOT, Yihong finished the TransCenter-DAUMOT part.
6 |
7 | [Guillaume Delorme*](https://team.inria.fr/robotlearn/team-members/guillaume-delorme/), [Yihong Xu*](https://team.inria.fr/robotlearn/team-members/yihong-xu/), [Luis G. Camara](https://team.inria.fr/robotlearn/team-members/luis-gomez-camara/), [Elisa Ricci](http://elisaricci.eu/), [Radu Horaud](https://team.inria.fr/perception/team-members/radu-patrice-horaud/), [Xavier Alameda-Pineda](http://xavirema.eu/)
8 |
9 |

10 |
11 |
12 | Sample qualitative results from different sequences of MOT20 comparing direct transfer (i.e. no adaptation) to DAUMOT. (a) shows the persons missed by
13 | direct transfer whereas (b) showcases over detections and (c,d) show person identity changes. The images illustrate the existence of large domain shifts between
14 | source (MOT17) and target (MOT20) when the model is not trained with DAUMOT. Green arrows point to the errors.
15 |
16 |
17 |
18 |
19 |
20 |
--------------------------------------------------------------------------------
/teaser_DAUMOT.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/yihongXU/DAUMOT/0dcb36d90c9683824e8f9540a176c733a0bb9300/teaser_DAUMOT.png
--------------------------------------------------------------------------------