Understanding Modern Computer Vision

A workshop walking through the evolution of modern computer vision — from Convolutional Neural Networks through to Vision Transformers.

What you will learn

• How CNNs detect and build visual features
• How Vision Transformers represent images as patch embeddings and use multi-head self-attention to develop global feature understanding
• The key differences, strengths and trade-offs between the two models

The workshop is delivered as a presentation with live visual demonstrations illustrating each model's internals and how they process images step by step.