[TAP-Vid] [TAPIR] [RoboTAP] [Blog Post] [BootsTAP] [TAPVid-3D]
tapir.mp4
Welcome to the official Google Deepmind repository for Tracking Any Point (TAP), home of the TAP-Vid and TAPVid-3D Datasets, our top-performing TAPIR model, and our RoboTAP extension.
- TAP-Vid is a benchmark for models that perform this task, with a collection of ground-truth points for both real and synthetic videos.
- TAPIR is a two-stage algorithm which employs two stages: 1) a matching stage, which independently locates a suitable candidate point match for the query point on every other frame, and (2) a refinement stage, which updates both the trajectory and query features based on local correlations. The resulting model is fast and surpasses all prior methods by a significant margin on the TAP-Vid benchmark.
- RoboTAP is a system which utilizes TAPIR point tracks to execute robotics manipulation tasks through efficient imitation in the real world. It also includes a dataset with ground-truth points annotated on real robotics manipulation videos.
- BootsTAP (or Bootstrapped Training for TAP) uses a large dataset of unlabeled, real-world video to improve tracking accuracy. Specifically, the model is trained to give consistent predictions across different spatial transformations and corruptions of the video, as well as different choices of the query points. We apply it to TAPIR to create BootsTAPIR, which is architecturally similar to TAPIR but substantially outperforms it on TAP-Vid.
- TAPVid-3D is a benchmark and set of metrics for models that perform the 3D point tracking task. The benchmark contains 1M+ computed ground-truth trajectories on 4,000+ real-world videos.
This repository contains the following:
- TAPIR / BootsTAPIR Demos for both online colab demo and offline real-time demo by cloning this repo
- TAP-Vid Benchmark for both evaluation dataset and evaluation metrics
- RoboTAP Benchmark for both evaluation dataset and point track based clustering code
- TAPVid-3D Benchmark for the evaluation metrics and sample evaluation code for the TAPVid-3D benchmark.
- Checkpoints for TAP-Net (the baseline presented in the TAP-Vid paper), TAPIR and BootsTAPIR pre-trained model weights in both Jax and PyTorch
- Instructions for training TAP-Net (the baseline presented in the TAP-Vid paper) and TAPIR on Kubric
The simplest way to run TAPIR / BootsTAPIR is to use our colab demos online. You can also clone this repo and run on your own hardware, including a real-time demo.
You can run colab demos to see how TAPIR works. You can also upload your own video and try point tracking with TAPIR. We provide a few colab demos:
Standard TAPIR: This is the most powerful TAPIR / BootsTAPIR model that runs on a whole video at once. We mainly report the results of this model in the paper.
Clone the repository:
git clone https://github.com/deepmind/tapnet.git
Switch to the project directory:
cd tapnet
Install the tapnet python package (and its requirements for running inference):
pip install .
Download the checkpoint
mkdir checkpoints
wget -P checkpoints https://storage.googleapis.com/dm-tapnet/causal_tapir_checkpoint.npyAdd current path (parent directory of where TapNet is installed)
to PYTHONPATH:
export PYTHONPATH=`(cd ../ && pwd)`:`pwd`:$PYTHONPATH
If you want to use CUDA, make sure you install the drivers and a version of JAX that's compatible with your CUDA and CUDNN versions. Refer to the jax manual to install the correct JAX version with CUDA.
You can then run a pretrained causal TAPIR model on a live camera and select points to track:
cd ..
python3 ./tapnet/live_demo.py \In our tests, we achieved ~17 fps on 480x480 images on a quadro RTX 4000 (a 2018 mobile GPU).
This repository hosts three separate but related benchmarks: TAP-Vid, its later extension RoboTAP, and TAPVid-3D.
tap-vid.mp4
TAP-Vid is a dataset of videos along with point tracks, either manually annotated or obtained from a simulator. The aim is to evaluate tracking of any trackable point on any solid physical surface. Algorithms receive a single query point on some frame, and must produce the rest of the track, i.e., including where that point has moved to (if visible), and whether it is visible, on every other frame. This requires point-level precision (unlike prior work on box and segment tracking) potentially on deformable surfaces (unlike structure from motion) over the long term (unlike optical flow) on potentially any object (i.e. class-agnostic, unlike prior class-specific keypoint tracking on humans).
More details on downloading, using, and evaluating on the TAP-Vid benchmark can be found in the corresponding README.
RoboTAP is a following work of TAP-Vid and TAPIR that demonstrates point tracking models are important for robotics.
The RoboTAP dataset follows the same annotation format as TAP-Vid, but is released as an addition to TAP-Vid. In terms of domain, RoboTAP dataset is mostly similar to TAP-Vid-RGB-Stacking, with a key difference that all robotics videos are real and manually annotated. Video sources and object categories are also more diversified. The benchmark dataset includes 265 videos, serving for evaluation purpose only. More details can be found in the TAP-Vid README. We also provide a 
TAPVid-3D is a dataset and benchmark for evaluating the task of long-range Tracking Any Point in 3D (TAP-3D).
The benchmark features 4,000+ real-world videos, along with their metric 3D position point trajectories. The dataset is contains three different video sources, and spans a variety of object types, motion patterns, and indoor and outdoor environments. This repository folder contains the code to download and generate these annotations and dataset samples to view. Be aware that it has a separate license from TAP-Vid.
More details on downloading, using, and evaluating on the TAPVid-3D benchmark can be found in the corresponding README.
In our storage datasets, (x, y) coordinates are typically in normalized raster coordinates: i.e., (0, 0) is the upper-left corner of the upper-left pixel, and (1, 1) is the lower-right corner of the lower-right pixel. Our code, however, immediately converts these to regular raster coordinates, matching the output of the Kubric reader: (0, 0) is the upper-left corner of the upper-left pixel, while (h, w) is the lower-right corner of the lower-right pixel, where h is the image height in pixels, and w is the respective width.
When working with 2D coordinates, we typically store them in the order (x, y). However, we typically work with 3D coordinates in the order (t, y, x), where y and x are raster coordinates as above, but t is in frame coordinates, i.e. 0 refers to the first frame, and 0.5 refers to halfway between the first and second frames. Please take care with this: one pixel error can make a difference according to our metrics.
tapnet/checkpoint/ must contain a file checkpoint.npy that's loadable using our NumpyFileCheckpointer. You can download checkpoints here, which should closely match the ones used in the paper.
Note: evaluation results in the table are reported under 256x256 inference resolution, but higher resolution can benefit results. For BootsTAPIR, we typically find the best results at 512x512 resolution, and for TAPIR, even higher resolutions than 512x512 can be beneficial.
| model | checkpoint | config | backbone | training resolution | DAVIS First (AJ) | DAVIS Strided (AJ) | Kinetics First (AJ) | RoboTAP First (AJ) |
|---|---|---|---|---|---|---|---|---|
| TAP-Net | Jax | tapnet_config.py | TSM-ResNet18 | 256x256 | 33.0% | 38.4% | 38.5% | 45.1% |
| TAPIR | Jax & PyTorch | tapir_config.py | ResNet18 | 256x256 | 58.5% | 63.3% | 50.0% | 59.6% |
| Online TAPIR | Jax | causal_tapir_config.py | ResNet18 | 256x256 | 56.2% | 58.3% | 51.2% | 59.1% |
| BootsTAPIR | Jax & PyTorch | tapir_bootstrap_config.py | ResNet18 + 4 Convs | 256x256 + 512x512 | 62.4% | 67.4% | 55.8% | 69.2% |
| Online BootsTAPIR | Jax & PyTorch | tapir_bootstrap_config.py | ResNet18 + 4 Convs | 256x256 + 512x512 | 59.7% | 61.2% | 55.1% | 69.1 |
We provide a train and eval framework for TAP-Net and TAPIR in the training directory; see the training README.
Please use the following bibtex entries to cite our work:
@article{doersch2022tap,
title={{TAP}-Vid: A Benchmark for Tracking Any Point in a Video},
author={Doersch, Carl and Gupta, Ankush and Markeeva, Larisa and Recasens, Adria and Smaira, Lucas and Aytar, Yusuf and Carreira, Joao and Zisserman, Andrew and Yang, Yi},
journal={Advances in Neural Information Processing Systems},
volume={35},
pages={13610--13626},
year={2022}
}
@inproceedings{doersch2023tapir,
title={{TAPIR}: Tracking any point with per-frame initialization and temporal refinement},
author={Doersch, Carl and Yang, Yi and Vecerik, Mel and Gokay, Dilara and Gupta, Ankush and Aytar, Yusuf and Carreira, Joao and Zisserman, Andrew},
booktitle={Proceedings of the IEEE/CVF International Conference on Computer Vision},
pages={10061--10072},
year={2023}
}
@article{vecerik2023robotap,
title={{RoboTAP}: Tracking arbitrary points for few-shot visual imitation},
author={Vecerik, Mel and Doersch, Carl and Yang, Yi and Davchev, Todor and Aytar, Yusuf and Zhou, Guangyao and Hadsell, Raia and Agapito, Lourdes and Scholz, Jon},
journal={International Conference on Robotics and Automation},
pages={5397--5403},
year={2024}
}
@article{doersch2024bootstap,
title={{BootsTAP}: Bootstrapped Training for Tracking-Any-Point},
author={Doersch, Carl and Luc, Pauline and Yang, Yi and Gokay, Dilara and Koppula, Skanda and Gupta, Ankush and Heyward, Joseph and Rocco, Ignacio and Goroshin, Ross and Carreira, Jo{\~a}o and Zisserman, Andrew},
journal={Asian Conference on Computer Vision},
year={2024}
}
@article{koppula2024tapvid,
title={{TAPVid}-{3D}: A Benchmark for Tracking Any Point in {3D}},
author={Koppula, Skanda and Rocco, Ignacio and Yang, Yi and Heyward, Joe and Carreira, Jo{\~a}o and Zisserman, Andrew and Brostow, Gabriel and Doersch, Carl},
journal={Advances in Neural Information Processing Systems},
year={2024}
}
Copyright 2022-2024 Google LLC
Software and other materials specific to the TAPVid-3D benchmark are covered by the license outlined in tapvid3d/LICENSE file.
All other software in this repository is licensed under the Apache License, Version 2.0 (Apache 2.0); you may not use this file except in compliance with the Apache 2.0 license. You may obtain a copy of the Apache 2.0 license at:
https://www.apache.org/licenses/LICENSE-2.0
All other non-software materials released here for the TAP-Vid datasets, i.e. the TAP-Vid annotations, as well as the RGB-Stacking videos and RoboTAP videos, are released under a Creative Commons BY license. You may obtain a copy of the CC-BY license at: https://creativecommons.org/licenses/by/4.0/legalcode .
The original source videos of DAVIS come from the val set, and are also licensed under creative commons licenses per their creators; see the DAVIS dataset for details. Kinetics videos are publicly available on YouTube, but subject to their own individual licenses. See the Kinetics dataset webpage for details.
Unless required by applicable law or agreed to in writing, all software and materials distributed here under the Apache 2.0 or CC-BY licenses are distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the licenses for the specific language governing permissions and limitations under those licenses.
This is not an official Google product.