VOT2014 stands for Visual Object Tracking Challenge 2014. The dataset includes 25 short sequences covering various objects in challenging backgrounds. Eight of the sequences are from the VOT2013 challenge. The newly added sequences show complementary objects and backgrounds, such as fish underwater or surfers riding big waves.

The sequences in this dataset were selected from a large library of sequences by clustering the visual features of objects and background, so that the 25 sequences were evenly sampled in the existing library of sequences.