New benchmark dataset tackles dense crowd counting challenges in Hajj footage

Researchers have introduced HAJJv2-CrowdCount, a new benchmark dataset for dense crowd counting specifically designed for Hajj video footage. This dataset addresses the challenges of steep camera angles, extensive occlusion, and high crowd density that typical models struggle with. Benchmarking three zero-shot counting methods, the study found that while SAM3Count performed best overall, a point-based counter (APGCC) proved more reliable in the densest, most occluded scenes, which are critical for Hajj crowd management. AI

IMPACT Provides a specialized benchmark for dense crowd counting, potentially improving AI's ability to manage large public gatherings.

RANK_REASON The cluster contains an academic paper introducing a new benchmark dataset and evaluating existing models. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• AlJawharh AlOtaibi
• APGCC
• HAJJv2
• HAJJv2-CrowdCount
• SAM3Count
• YOLO-World

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →