Post by Ali Abdullah

Dusting off the old LinkedIn profile because my summer vacation has officially hit the "I have absolutely nothing to do" phase. 😅 So, what better time to start uploading all the cool stuff I've built? Brace yourselves, I’m going to be posting a bunch of my projects over the summer! Kicking things off with SmartCrowdDetector—my 3rd-semester Software Engineering project! 🚀 This wasn't just a regular classroom assignment. To get this project, our team literally had to go down to NSTP at NUST, pitch ourselves, and convince Zaytrics to trust us with a real-world problem. And honestly? We delivered. So, what did we build? SmartCrowdDetector is a computer vision pipeline that performs real-time crowd tracking and generates dynamic density heatmaps to analyze congested areas. Time for the technical LARP 🤓: We built the core detection engine using YOLOv8, leveraging .onnx and .engine (TensorRT) exports to optimize inference speeds so it could run blazingly fast. The Python backend orchestrates custom video stream handlers, runs the bounding-box logic through a density mapping algorithm, and serves the analytics via a clean web interface. To top it off, we containerized the entire environment using Docker so it's deploy-ready out of the box. Massive shoutout to my amazing team who survived this grind with me: Muhammad Omer Irfan, Mutee Ullah Akhtar Mughal, and Muhammad Musa Nadeem. And a huge thanks to Zaytrics for taking a chance on a bunch of 3rd-semester students! Check out the demo video below! 👇 All the code is fully open-source and documented here: https://lnkd.in/dzjxupWE Stay tuned for more random projects as I continue to be bored this summer! ☀️ #ComputerVision #YOLOv8 #SoftwareEngineering #Python #MachineLearning #NUST #NSTP #Docker #OpenSource

Video Content