VOT2014 (Visual Object Tracking Challenge 2014)

该数据集包含 25 个短序列，显示具有挑战性的背景中的各种对象。八个序列来自 VOT2013 挑战（螺栓、自行车、大卫、潜水、体操、手、遮阳伞、女人）。新序列显示互补的对象和背景，例如水下的鱼或冲浪者在大浪中骑行。使用基于聚类对象和背景的视觉特征的方法从大量序列中选择序列，以便这 25 个序列在现有池中均匀采样。This dataset comprises 25 short sequences showcasing various objects against challenging backgrounds. Eight of the sequences are sourced from the VOT2013 challenge, including bolt, bicycle, David, diving, gymnastics, hand, umbrella, and woman. The remaining new sequences feature complementary objects and backgrounds, such as underwater fish or surfers riding large waves. Sequences were selected from a large candidate pool using a method that clusters visual features of both objects and backgrounds, ensuring these 25 sequences are uniformly sampled from the pool.

搜集汇总

背景与挑战

背景概述

VOT2014是一个视觉目标跟踪数据集，包含25个短序列，展示了在复杂背景下的多种对象，部分序列继承自VOT2013。这些序列通过视觉特征聚类方法选取，以确保在现有池中均匀分布，适用于目标跟踪和检测任务。