Computer Vision — ZeroCourse

reading · we frame, you read MIT or the canonical taught · we author, no canonical fits ↺ spirals back to earlier lessons

Course locked

Complete Neural Networks and Deep Learning Fundamentals first.

This course unlocks once you've finished its prerequisite. Open prerequisite →

Prerequisites

Neural Networks and Deep Learning Fundamentals not yet

Complete prerequisites first to enroll.

reading · 35 min + ~240 min canonical

Image Fundamentals and Classical Computer Vision

Before deep learning taught machines to see, an entire field spent decades teaching them to see the hard way — by hand-crafting the algorithms. This lesson is that foundation, and…

primary Stanford CS231N — Convolutional Neural Networks for Visual Recognition (YouTube)

preview →

reading · 35 min + ~240 min canonical L1

Image Classification: The ImageNet Moment and Modern Architectures

Image classification — "what is in this picture?" — is the foundational computer-vision task, the one that started the deep-learning revolution and the one every other vision task…

primary Stanford CS231N — Convolutional Neural Networks for Visual Recognition (YouTube)

after L1 locked

reading · 35 min + ~240 min canonical L1

Object Detection: Finding What and Where

Image classification (last lesson) answers "what is in this picture?" — but that's not enough for most real applications. A self-driving car doesn't just need to know "there's a p…

primary Stanford CS231N — Convolutional Neural Networks for Visual Recognition (YouTube)

after L2 locked

reading · 35 min + ~240 min canonical L1

Image Segmentation: Pixel-Level Understanding

Classification said what (one label per image). Detection said what and where (a box per object). Segmentation goes all the way: it labels every single pixel — the finest-grained …

primary Stanford CS231N — Convolutional Neural Networks for Visual Recognition (YouTube)

after L3 locked

reading · 40 min + ~240 min canonical L1

Vision Transformers, Multimodal Models, and Real-World Applications

Two developments reshaped computer vision after the CNN era, and this lesson covers both — then closes with the honest reckoning that must accompany any powerful technology. The f…

primary Stanford CS231N — Convolutional Neural Networks for Visual Recognition (YouTube)

after L4 locked

exercise reading · 15 min L1

Exercises: Easy (Warm-up)

- [ ] Implement convolution from scratch in Python — 2D convolution with various kernels (Gaussian blur, Sobel edge detection) - [ ] Image classification with transfer learning — …

after L5 locked

exercise reading · 30 min L1

Exercises: Medium (Core)

- [ ] Kaggle: Dogs vs Cats — Binary image classification with CNN, data augmentation, transfer learning - [ ] Implement data augmentation pipeline — Random crop, flip, rotation, c…

after L6 locked

exercise reading · 45 min L1

Exercises: Hard (Stretch)

- [ ] Build a real-time object detector — YOLO running on webcam feed, draw bounding boxes, label objects - [ ] Implement image style transfer — Gram matrix matching with a pre-tr…

after L7 locked

exercise reading · 60 min L1

Mini-Project

- [ ] Build a plant disease identifier — Collect/download plant leaf images, train a CNN (ResNet or EfficientNet), build a Gradio web interface for uploading and classifying; eval…

after L8 locked

L10

project reading · 360 min L1

Project: Image Classifier with Transfer Learning

Fine-tune a pre-trained ResNet-50 on a custom dataset (collect 10 classes, 100 images each from the internet). Use PyTorch, data augmentation, and a cosine annealing learning rate…

after L9 locked

L11

project reading · 300 min L1

Project: Object Detection with YOLO

Use YOLOv8 (Ultralytics) to train a custom object detector on a small labeled dataset (COCO subset or your own). Evaluate mAP, visualize predictions, and run inference on test ima…

after L10 locked

L12

assessment reading · 56 min L1

Self-Assessment: Computer Vision

- [ ] How does a convolutional layer work? Explain kernels, stride, padding, and output size calculation. - [ ] What is ResNet? How do skip connections solve the vanishing gradien…

after L11 locked

12 lessons. Read in order; spiral back when you need to. By the end you'll have used the core ideas twice — once on the abstract, once on something you'll meet at work next week.