Oumar Ibrahim is a senior computer engineering student at the University of Sharjah (3.91 CGPA) researching lightweight, multi-modal sensing for Driver Monitoring Systems, including a conference paper benchmarking RGB detectors and an eight-model NIR study, and for MANAR, a Search-and-Rescue drone system that fuses RGB, NIR, thermal, RF, and mmWave sensing via Selective State Space Models like Mamba.
| Repository | Focus Area | Description |
|---|---|---|
| class-watcher-uos | Public Utility / Automation | Real-time course seat monitor for University of Sharjah registration with instant phone push, Windows toasts, email alerts, and a standalone 1-click installer. |
| lightweight-driver-behavior-detection | Edge AI / Computer Vision | Benchmark of lightweight object detection architectures for real-time driver state monitoring under a strict subject-disjoint evaluation protocol. |
| mmwave-pose-physics | Radar / Signal Processing | Controlled evaluation of accuracy, calibration, and selective prediction for lightweight (<5M parameter) mmWave human-pose estimation under signal degradation. |
| manar-search-rescue-drone | Robotics / Sensor Fusion | Supervised-autonomy multisensor search-and-rescue drone combining RGB/IR/thermal vision, FMCW radar, passive RF, audio, and temporal sensor fusion. |
| macroblock-visual-automation | Desktop Automation / Systems | Visual Windows desktop automation environment allowing users to construct complex system workflows by snapping functional blocks together, with a global abort hotkey. |
If you are working on edge AI, radar perception, or applied signal processing, feel free to connect on LinkedIn.
