You Only Look Once: The Journey of Open Source Giving AI Eyes

Object detection is not only for AI labs anymore — with open-source tools for computer vision, you can build smart applications, real-time demos, and AI-powered projects, all while fully understanding the implications of how it works.

In this talk, I want to take you along my journey. I went from confused and disconnected to building my own real-time object detection demo using ONLY free and open source tools . I'll explain what YOLO (You Only Look Once) is, in simple language. It's not all just scary math. It's a tool; a thing that someone built for you to use, learn from, improve on and build upon.

This is not just technology. It is an invitation. And it's open to anyone.

We'll cover:

Why Open-Source AI is Beneficial for Beginners

• It's free, flexible, and runs on hardware you already own (or free cloud GPUs!).

• It's built by a community that wants you to learn and give back, not just consume.

The FOSS AI Toolbox for Computer Vision

• YOLO (You Only Look Once) - The super fast, accurate object recognition algorithm that you can use for free(ultralytics).

• Google Colab - Your live action, GPU powered environment for running AI models in the browser, no setup needed. \

• Hugging Face & Github - Where you can find the pre-trained models, datasets, and code to get you started.

• Roboflow - For organizing and preparing your custom image datasets.

Live Demo (Live)

• Use google collab notebook

• Use a pre-written trained model → upload a picture (from website or local device) →objects will be detected by YOLO algorithm using ultralytics tool.

• Show how fast AI models can be made using Open Source tools and libraries only.

What You Can Create

• Hackathon prototypes that are astonishingly quick and astonishingly good.

• The building blocks of more sophisticated applications in areas such as robotics, accessibility, and data analysis.

How to Contribute Back (You Don't Need Expert-Level Skills!)

• You don't need to invent a new algorithm to contribute back. I can show you how I got started by:

• Improving documentation and tutorials for other beginners.

• Testing models and reporting bugs.

• A straightforward, clearly-articulated description of how object detection does work.

• A practical roadmap to free tools and resources to kick off their own simple AI project this weekend.

• The belief that they have value and can use and contribute to open-source AI, wherever they currently are at skill-wise.

• The belief to view AI not as magic, but as a tool they can build with and is powerful, yet accessible.