PANDA: A Gigapixel-level Human-centric Video Dataset
This dataset addresses the problem of analyzing human behaviors and interactions in large-scale real-world scenes for the AI and praxeology communities, but it is incremental as it primarily provides new data rather than a novel method.
The authors introduced PANDA, a gigapixel-level human-centric video dataset for large-scale visual analysis, capturing real-world scenes with wide field-of-view and high resolution, and benchmarked tasks like human detection and tracking, showing that existing methods struggle with accuracy and efficiency due to scale variations and occlusions.
We present PANDA, the first gigaPixel-level humAN-centric viDeo dAtaset, for large-scale, long-term, and multi-object visual analysis. The videos in PANDA were captured by a gigapixel camera and cover real-world scenes with both wide field-of-view (~1 square kilometer area) and high-resolution details (~gigapixel-level/frame). The scenes may contain 4k head counts with over 100x scale variation. PANDA provides enriched and hierarchical ground-truth annotations, including 15,974.6k bounding boxes, 111.8k fine-grained attribute labels, 12.7k trajectories, 2.2k groups and 2.9k interactions. We benchmark the human detection and tracking tasks. Due to the vast variance of pedestrian pose, scale, occlusion and trajectory, existing approaches are challenged by both accuracy and efficiency. Given the uniqueness of PANDA with both wide FoV and high resolution, a new task of interaction-aware group detection is introduced. We design a 'global-to-local zoom-in' framework, where global trajectories and local interactions are simultaneously encoded, yielding promising results. We believe PANDA will contribute to the community of artificial intelligence and praxeology by understanding human behaviors and interactions in large-scale real-world scenes. PANDA Website: http://www.panda-dataset.com.