Rudransh Dixit@rudransh_dixit

Security teams often monitor many camera feeds at once, making it easy to miss a fall, fight, accident or rule violation until someone reviews the footage later. OneEye helps operators move from reactive footage review to earlier, evidence-based response. OneEye is a web dashboard connected to live camera feeds and five computer-vision pipelines: fight detection, fallen-person detection, accident detection, garbage detection and mobile-phone detection. OpenCV captures frames; the models analyse them; and a FastAPI service records events and provides camera, alert, analytics and settings APIs. Operators can view streams, review time-stamped events and supporting images or clips, inspect model confidence, and adjust alert thresholds. The primary users are security teams, facility managers, campus administrators and other staff responsible for monitored spaces. A user starts a configured camera, watches the live feed and checks the event list when an alert appears. This brings several safety and operational checks into one workflow instead of requiring a person to watch every feed continuously. OneEye’s distinctive approach combines five event types, shared camera processing, evidence-backed alerts and a single dashboard. Consecutive-frame confirmation and cooldowns help reduce repeated alerts. It is a working local prototype whose models can be evaluated and improved individually for each deployment setting.

💻 Online Hackathon | ⏳ Build: 48 hours | 📅 September 25–27, 2026 The official IBM Bob 2.0 hackathon. Build what's next in AI-assisted development, alongside the developers and experts shaping it. ⭐ Get help from experts throughout the event 🤝 Join a team or build solo 💰 Prize pool: $12,000 🧑🏻‍💻 Register before the Kick-Off Stream to start building with Bob