Computer Vision: Foundations & Applications
How a computer turns a grid of brightness numbers into 'that's a face,' and why a camera is not an eye.
AI tutor is turned off for this class
Use the CS131 lectures, notes, and assignments below to keep learning.
Lesson overview
How a computer turns a grid of brightness numbers into 'that's a face,' and why a camera is not an eye.
Teacher script · 45 min
- 0–5
Hook
To a computer, a photo is a giant spreadsheet where each cell holds how bright that dot is. There's no 'cat' in there, just numbers. All of computer vision is finding patterns in those numbers that reliably mean 'cat,' 'edge,' or 'face.'
- 5–15
Explore
Students do the activity in pairs: Print a photo, lay tracing paper over it, and trace only where light meets dark.
- 15–30
Explain
A face looks different up close, far away, tilted, or in shadow. The goal is to find 'features' (corners, textures, patterns) that stay recognizable through those changes. Classic vision hand-designs these; modern vision learns them with convolutional networks. Either way, the win is descriptions that are stable under transformation.
- 30–40
Connect to the summit
Read aloud the "From the summit" section below. Ask students which ideas from today they recognize in it.
- 40–45
Check
Run the formative check below. Anyone who can explain a key term in their own words has it.
Student activity
Print a photo, lay tracing paper over it, and trace only where light meets dark.
Slides
Formative check
- 1.In your own words, what is "Pixel"? (Looking for: One dot of an image, stored as a number (or three) for its brightness/color.)
- 2.In your own words, what is "Convolution"? (Looking for: Sliding a small pattern over an image to detect where that pattern appears.)
- 3.In your own words, what is "Edge detection"? (Looking for: Finding sharp brightness changes, which usually mark the boundaries of objects.)
Carry-away concepts
- Pixel
- One dot of an image, stored as a number (or three) for its brightness/color.
- Convolution
- Sliding a small pattern over an image to detect where that pattern appears.
- Edge detection
- Finding sharp brightness changes, which usually mark the boundaries of objects.
- Feature
- A compact, reliable description of a piece of an image that survives changes in view.
From the summit · the Stanford source
You go from raw pixels through edges, features, and segmentation to recognition, mixing classical image processing with modern learning.
This module is based on a Stanford course.