AVisT offers a dedicated Visual Object Tracking dataset that covers a variety of adverse scenarios highly relevant to real-world applications. Importantly, it poses additional challenges to the tracker design due to adverse visibility.

• AVisT covers a wide range of 18 diverse scenarios including rain, fog, hurricane, fire, sun glare, low-light, archival videos, fast motion, distractor objects, occlusion, snow, sandstorm, tornado, smoke, splashing water, camouflage, small objects and deformation. These diverse scenarios are broadly categorized into five attributes: weather conditions, obstruction effects, imaging effects, target effects and camouflage.

• AVisT dataset comprises 120 challenging videos from YouTube under the Creative Commons license with a total of approximately 80k (79653) annotated frames. The frame-rates of these videos ranges from 24 to 30 frames per second (fps) and the average sequence length is 664 frames (i.e., 22.2 seconds with 30 fps). The shortest sequence in the dataset has 99 frames (3.3 seconds with 30 fps), while the longest one has 3113 frames (103.7 seconds with 30 fps).

Obstruction Effects

Imaging Effects

Target Effects

Camouflage