APIDIS

Data

Approximatively 2 hours of videos were captured from 7 viewpoints during a professional basketball game.

Raw data

The cameras were recording at almost 22 fps in average, with a resolution of 1600x1200 pixels. The video files are available in their native format, i.e. one motion jpeg file (~300 MB) per minute per camera.

Pseudo-synchronised video

A pseudo synchronised dataset is provided by resampling the raw videos at 25 fps (using the closest available frame) with a resolution of 800x600 pixels. The video files are available as one MPEG-4 file (between 28 MB and 56 MB) per minute per camera. For optimal timestamps accuracy, the original dataset should be prefered.

Annotations

Timestamps

All timestamps are expressed in seconds since Epoch when provided as integers. When provided in a human readable format, e.g. in filenames, they follow the ISO 8601 date/time syntax.

Calibration

All cameras are Arecont Vision AV2100M IP cameras (datasheet). The fish-eye lenses used for the top view cameras are Fujinon FE185C086HA-1 lenses (datasheet). Each camera was indivually calibrated into a common world coordinate system.

Additional annotations

Several manually annotated objects are provided (see NEM'08 summit paper): - Basket ball events like ball possession periods, throws, violations ; for the whole basketball game - Ball, Players and referees position ; for one minute of the game.

Annotated events and salient-objects are recorded into two kinds of XML files whose syntax is described in the XML Schema Definition (xsd) files apidis-annotation-ver23.xsd and apidis-salientobj-ver1.xsd. A simplified structural diagram of event xml files is illustrated in event-xml-simple.png. The tags for describing the detected objects and their properties are illlustrated in salient-obj-xml.png.

Ball was annotated in the 2D images on every camera where it was visible. An approximate 3D localization is inferred for the pseudo-synchronized dataset.

Terms of use:

This dataset is available for non-commercial research in video signal processing only. We kindly ask you to mention the APIDIS project when using this dataset (in publications, video demonstrations...).

Acknowledgements

We would like to thank Jean-François Prior (Dexia Namur basket ball team), Philippe Delmulle Declercq Stortbeton Waregem basket ball team) and the city of Namur for their authorisations and technical help collecting this dataset.