Football Dataset


The image sequences (and corresponding images) for the football dataset may be obtained from Or alternatively access the FTP site anonymously in the usual way, i.e.:

 Name: ftp
 ftp> binary
 ftp> prompt
 ftp> cd VS-PETS
 ftp> mget Frame0*

Note that it is crucial to employ "binary" transfer.

Note: there appears to be some problems accessing the ftp site using Netscape. If you are having problems, please try using Internet Explorer instead, or access via direct ftp as shown above.

Important instructions are given at the bottom of this page on processing the datasets - please read these carefully.

The datasets consists of football players moving around a pitch.  The task is to determine the location of each player on the pitch, for each frame of the sequence.

For each player, the bounding box (with origin bottom left) in pixels should be determined.  The posiiton of the player is defined as the middle bottom of the bounding box (in pixels).

The datasets are available under here

The coordinate space (plan of camera locations) is shown below.  Details on camera calibration are available in the ZIP file located here.

The entire camera 3 training set (zipped file) can be downloaded here.
The entire camera 4 training set (zipped file) can be downloaded here.
The entire camera 5 training set can be downloaded here.

Note: The training data is intended for developing background models of the scene.  The sequences has been specifically chosen to contain the least amount of player action.

The entire camera 3 test set (zipped file) can be downloaded here.
The entire camera 4 test set (zipped file) can be downloaded here.
The entire camera 5 test set can be downloaded here.

The annotation (ground truth) for camera view 3 is available here, with details of the format available here.  An AVI file of the ground truth for camera view 3 is also available here.

The annotation for camera views 4 & 5 will be available soon.

You can also download all files under one directory using wget i.e.
wget -m
will download all camera 3 test images to the current directory.  Please see  for more details.

If you are having problems, please access via direct ftp as shown above.

There are two directories - one for training, and one for testing.  Results should be presented only on the test data - see below for more detailed instructions.

Example images for two (synchronised) camera views


Camera views 3, 4 and 5


Camera views 3, 4 and 5

Notes on datasets:

The tracking/counting can be performed on the entire test sequence (for either or both camera views), or a portion of it. The images may be converted to any other format as appropriate, e.g. subsampled, or converted to monochrome. All results reported in the paper should clearly indicate which part of the test data is used, ideally with reference to frame numbers where appropriate, e.g. CAMERA 3, TEST DATASET, FRAMES 1-100.

There is no requirement to report results on the whole test sequence (camera 3 and/or 4), however you are encouraged to test your algorithms on as much of the test data as possible.

Please ensure when quoting training or test sequences in the paper, that they match the datasets on this website.

The sequences have been provided by the consortium of Project IST INMOVE (IST-2001-37422)


The results that you report in your paper submission

The paper that you submit may be based on previously published tracking methods/algorithms (including papers submitted to the main ICCV conference).  The importance is that your paper MUST report results using the datasets supplied.

If you have any queries please email