We propose the first higher frame rate video dataset
(called Need for Speed - NfS) and benchmark for
visual object tracking. The dataset consists of 100
videos (380K frames) captured with now commonly
available higher frame rate (240 FPS) cameras from
real world scenarios. All frames are annotated with
axis aligned bounding boxes and all sequences are
manually labelled with nine visual attributes - such
as occlusion, fast motion, background clutter,
etc. Our benchmark provides an extensive evaluation
of many recent and state-of-the-art trackers on
higher frame rate sequences. We ranked each of these
trackers according to their tracking accuracy and
real-time performance. One of our surprising
conclusions is that at higher frame rates, simple
trackers such as correlation filters outperform
complex methods based on deep networks. This
suggests that for practical applications (such as in
robotics or embedded vision), one needs to carefully
tradeoff bandwidth constraints associated with
higher frame rate acquisition, computational costs
of real-time analysis, and the required application
accuracy. Our dataset and benchmark allows for the
first time (to our knowledge) systematic exploration
of such issues.
Download
To download the dataset, *nix users can run the following command:
Video previews are available on the embedded YouTube playlist.
About the Dataset
We provide each video sequence as a seperate .zip file, each containing bounding box annotations and frames (as JPEG files) for the two scenarios described in our paper:
240 FPS capture, and 30 FPS with synthesized motion blur (generated with Adobe After Effects).
The annotation files are plain text, and were generated with Vatic.
We also make available all results from the trackers evaluated in our paper. (MD5)
The full tracking suite used to generate our results, and interface with our dataset is also available here. Each individual tracker is the property of the respective
authors, and we make no claim of ownership to these works. By making this tracking suite available, we allow our results to be replicated by interested parties. [COMING SOON]
Publications
The following publications are associated with our dataset.
H. Kiani Galoogahi, A. Fagg, C. Huang, D. Ramanan, S.Lucey. Need for Speed: A Benchmark for Higher Frame Rate Object Tracking, 2017,
arXiv preprint arXiv:1703.05884 - [PDF]
If you use the NfS dataset, we ask that you cite us.
