Roadmap of MMTracking about mmtracking HOT 17 OPEN

Hi @SilvioGiancola @agirbau, thanks for your interest!

We are discussing a development plan in which we tend to support 2-3 new methods for each task and support the other popular datasets.

For the methods, the Siam-Like and some recent SOTA methods (not decided which one yet) are the first priority for SOT @SilvioGiancola. As @agirbau suggested, We do have a plan to support CenterTrack but need to wait for the CenterNet in mmdetection.

For the datasets, we will support the datasets @SilvioGiancola mentioned, and also support some other large scale MOT benchmarks.

However, considering the limited bandwidth, we cannot clearly state the release schedule of these updates. If you are also interested in supporting these features, we will really appreciate your pull requests. : )

from mmtracking.

What about HOTA metrics?

from mmtracking.

Hello! I think CenterTrack would be a great incorporation to the multiple object tracking.

from mmtracking.

VID feature request like mmdetection toolmmdetection/tools/analysis_tools(eg. visualize dataset's ground truth, output confusion matrix).

from mmtracking.

Pose track is also a good to-do as a new task, baseline methods include PoseFlow[1]. The corresponding datasets are PoseTrack[2] and HiEve[3].

[1] "Pose Flow: Efficient online pose tracking"
[2] "Posetrack: A benchmark for human pose estimation and tracking"
[3] "Human in events: A large-scale benchmark for human-centric video analysis in complex events"

An existing codebase Alphapose may be a good reference for the practice, which focuses on pose estimation while leaving tracking modules not actively maintained.

from mmtracking.

Thanks for your hard work on maintaining this wonderful repo.
It would be nice to include recently released large-vocabulary tracking data TAO as well.

from mmtracking.

When will release the CenterTrack in your plan?

from mmtracking.

Hi, I'd like to suggest StrongSORT MOT algorithm #541

from mmtracking.

I am considering to make support to some features in the short future. Let me know if some are already in process or your vote for their priority.

1. Support of DanceTrack[0] dataset.( #543)
2. Support of OC-SORT[6] for multi-object tracking. (#545 )
3. multi-class multi-object tracking, which is already supported in Youtube-VIS but still absent on MOT task. (#548)
4. transformer-based detector, e.g. DETR[1], and tracker, e.g. TransTrack[2], Trackformer[3].
5. more MOT datasets, such as KITTI[4].
6. make BYTE[5] and OC-SORT optional flags in other KalmanFilter-based tracking methods.

[0] "DanceTrack: Multi-Object Tracking in Uniform Appearance and Diverse Motion"
[1] "End-to-end object detection with transformers"
[2] "Transtrack: Multiple object tracking with transformer"
[3] "Trackformer: Multi-object tracking with transformers"
[4] "Vision meets robotics: The kitti dataset"
[5] "ByteTrack: Multi-Object Tracking by Associating Every Detection Box"
[6] "Observation-Centric SORT: Rethinking SORT for Robust Multi-Object Tracking"

from mmtracking.

Thank you for the amazing work! Do you have any plan on releasing more baselines on Single Object Tracking (Siam-like, ATOM, DiMP, OCEAN, KYS, LWL) as well as more dataset (OTB, VOT, TrackingNet, UAV123, NFS)?

from mmtracking.

Just wanted to add, it would also be cool, If a multi object tracking and segmentation baseline could be added like
(TrackRCNN)
https://github.com/VisualComputingInstitute/TrackR-CNN
for example.
As many benchmarks and workshops are introducing datasets and problems specifically in this domain such as MOTS CHALLENGE, KITTI MOTS, BDD100K MOTS etc.

And again, really appreciate this amazing work, for putting everything in such a modular perspective with efficiency.

from mmtracking.

Agree with @Cavaradossi more analysis tools like mmdetection would be great help. Currently experiencing problems in creating one of my own for plotting the confusion matrix.

from mmtracking.

Hello! We are the authors of this paper. We implemented our method fully based on MMTracking 0.x and experimented some more datasets like UVO v0.5 and OVIS. I see previous reply mentioned transformer-base detection/tracking which is what we have. Would be interested in creating a PR if you think it is a good idea to include our model. Thanks!

from mmtracking.

Hello! We are the authors of this paper. We implemented our method fully based on MMTracking 0.x and experimented some more datasets like UVO v0.5 and OVIS. I see previous reply mentioned transformer-base detection/tracking which is what we have. Would be interested in creating a PR if you think it is a good idea to include our model. Thanks!

Hi~, We are glad you're willing to create a pr to implement your model in MMTracking.

from mmtracking.

Hi! Could I ask if you have plans to include Bot-SORT as part of mmtracking?

from mmtracking.

+1 to getting BoT-SORT implemented

from mmtracking.

MOTRv2

https://github.com/megvii-research/MOTRv2

from mmtracking.