Rajat(Self)

CrowdGuard AI — From Crowd Intelligence to Human Safety The inspiration behind CrowdGuard AI comes from the heartbreaking Kolkata pandal tragedy. Reading this NDTV report made me think: *Can AI help identify crowd risks before they turn into disasters?* Link: https://www.ndtv.com/india-news/dont-come-near-the-pandal-its-collapsing-videos-point-to-how-kolkata-stampede-broke-out-12202672 That question became - CrowdGuard AI, an AI-powered crowd-safety decision-support platform designed to help event organizers and safety personnel monitor crowd conditions, identify potential risks, and make informed decisions. The project showcases Meta AI models, combining SAM 3.1 visual intelligence for people-counting workflows with Muse Spark reasoning capabilities to interpret crowd observations and explain potential safety concerns. Beyond traditional surveillance, CrowdGuard AI aims to transform visual observations into actionable safety intelligence through crowd monitoring, forecasting, and incident tracking. Its potential applications span religious gatherings, concerts, stadiums, railway stations, and public events worldwide. My vision is to demonstrate how powerful AI models can contribute to a safer world.