This page contains links to data used in the following publication: Lebeda, Hadfield, Matas, Bowden: Texture-Independent Long-Term Tracking Using Virtual Corners. In IEEE Transactions on Image Processing, 2016. Please cite this publication if using any of the data presented. Please, let me know if you find anything missing on this page.
Short-term tracking dataset
- New texture-less sequences Mug and Page; existing sequences Dog and Dudek: CMP::Karel Lebeda::ACCV2012 (or backup).
- New transparent sequence Spaceship: spaceship.zip.
Long-term tracking dataset
- New traffic sequences Nissan and LiverRun from the ytLongTrack dataset: CMP::Data::ytLongTrack.
- Existing sequences CarChase, Panda and Volkswagen from the TLD dataset: CVSSP::Zdenek Kalal::TLD dataset.
Benchmarks
- Visual Object Tracking challenge: 2013, 2014, 2015.
- Visual Tracker Benchmark: v1.0 and v1.1 (or 2.0).