← All learning paths
Learning pathBeginner~12 hours

Computer Vision Foundations

A structured path from light and pixels to geometry, classical features, and modern deep networks — with checkpoints and exercises at every step.

What you'll learn

  • Imaging
  • Geometry
  • Features
  • Deep learning
  • Deployment

Your progress

0 / 6 lessons reached

Lessons in this path

Work top to bottom within each module, or jump in from the table of contents on each lesson page.

Module 1. Imaging & digital images

How cameras turn light into arrays of numbers, and how we preprocess those arrays for algorithms.

  1. Lesson 155 min

    Light, sensors, and the imaging pipeline

    Radiance, photons, sensor noise, demosaicing, gamma, and the end-to-end path from scene to stored image.

  2. Lesson 260 min

    Pixels, convolution, and edges

    Grids, channels, linear filters, separable kernels, gradients, and building intuition before neural networks.

Module 2. Geometry & correspondence

Pinhole cameras, projection, epipolar geometry, and how 2D images relate to 3D structure.

  1. Lesson 365 min

    Camera models and projection

    Intrinsic and extrinsic parameters, the pinhole model, lens distortion, and projecting 3D points into pixels.

  2. Lesson 470 min

    Features, matching, and robust estimation

    Corners, descriptors, putative matches, RANSAC, homographies, and when planar models break.

Module 3. Learning-based vision

Convolutional networks for classification, detection, and segmentation — plus practical constraints on mobile and edge.

  1. Lesson 575 min

    Convolutional networks for images

    From fully-connected layers to conv blocks, receptive fields, pooling, and training objectives.

  2. Lesson 670 min

    Detection, segmentation, and on-device trade-offs

    Anchors vs queries, instance vs semantic segmentation, latency, memory, and quantization at a high level.