Bryan Anton, Eva Anton and Eric Feron, King Abdullah University of Science and Technology, Saudi Arabia

Crowds in general, and dense crowds in particular, create complex problems related to organization, logistics, security, etc. These problems appear in various contexts, including stadiums during sporting or musical events, pilgrimages (all religions combined) and even in certain megacities road intersections. This review explores the evolution of artificial intelligence (AI) techniques in crowd analysis, emphasizing tracking and counting methodologies. It delves into traditional and deep learn-ing approaches, evaluates benchmark datasets, discusses current limitations, and highlights potential improvements. A particular focus is given to applications during the Hajj pilgrimage in Mecca, illustrating the challenges and advancements in managing massive crowd gatherings.

Dense crowds, Hajj, AI, Deep learning, Datasets