BLVD is a large-scale 5D semantic dataset built for autonomous driving, which supports dynamic 4D (3D+temporal) tracking, 5D (4D+interactive) interactive event recognition, and intention prediction tasks.

The dataset contains 654 high-resolution video clips with a total of 120,000 frames, including 249,129 3D annotated frames, 4,902 independent frames for tracking (with a total length of 214,922 points), 6,004 valid segments for 5D interactive event recognition, and 4,900 frames for 5D intent prediction.