From Pixels to Patterns: A Pragmatist’s Guide to Computer Vision

Everyone is talking about ChatGPT and LLMs these days, but there is much more to AI than that! Computer Vision (CV) is one of the most powerful and transformative fields in AI, yet it can often feel like an intimidating black box of complex math and specialized hardware. But what if you could understand and even start building computer vision applications with the skills you already have as a developer? This talk demystifies the world of pixels and patterns, offering a pragmatic guide for the curious coder.

We will journey through the core concepts of computer vision, starting from the fundamental question: how does a computer “see”? We’ll explore key tasks like image classification, object detection, segmentation, and model deployment, breaking them down into understandable components. Using practical, python-centric examples, we’ll see how pre-trained models can be leveraged to solve real-world problems without needing a PhD in mathematics.

Attendees will leave this session with a clear mental model of the computer vision landscape, an understanding of the common tools and techniques (like OpenCV, ONNX, TensorRT, SAM-3, and more), and a practical framework for identifying problems that can be solved with CV.

You’ll gain the confidence to start your own computer vision project, whether it’s for a hobby, a hackathon, or the next big feature at your company. This isn’t about abstract theory; it’s about empowering you to turn images and videos into actionable intelligence.