Learn Computer Vision

Learn Computer Vision, with Hany Farid

These lectures introduce the theoretical and practical aspects of computer vision from the basics of the image formation process in digital cameras, through basic image processing, space/frequency representations, and techniques for image analysis, recognition, and understanding.

See also Learn to Code in Python

Computer Vision

Introduction
- welcome [video]
Image formation
- pinhole camera
  - images [video]
  - camera obscura [video]
  - perspective projection, 2-D [video]
  - perspective projection, 2-D variant [video]
  - perspective projection, 2-D generalized [video]
  - perspective projection, 3-D [video]
- lenses
  - thin lens [video]
  - depth of focus [video]
  - exposure [video]
- pixels
  - color filter array [video]
  - displays [video]
- artifacts
  - chromatic aberrations and noise [video]
  - lens distortion [video]
  - JPEG compression [video]
- summary [video]
Image filtering
- convolution
  - discrete-time signals and systems [video]
  - linear time-invariant systems, 1-D [video]
  - convolution, 1-D [video]
  - linear time-invariant systems, 2-D [video]
  - convolution, 2-D [video]
  - separable convolution [video]
- space and frequency
  - canonical basis [video]
  - Fourier, 1-D [video]
  - complex exponential [video]
  - Fourier, 2-D [video]
  - continuous to discrete sampling, space [video]
  - continuous to discrete sampling, frequency [video]
  - discrete to discrete sampling [video]
- pyramids
  - Gaussian pyramid [video]
  - Laplacian pyramid [video]
- features
  - edges [video]
  - edge detection [video]
  - line detection [video]
  - histogram of gradients (HOG) [video]
Image analysis
- motion
  - differential motion [video]
  - differential motion, implementation [video]
  - feature tracking [video]
  - feature tracking, implementation [video]
- stereo
  - depth from stereo [video]
  - epipolar constraints [video]
- homography
  - planar homography [video]
  - planar homography, application [video]
Image understanding
- overview [video]
- supervised learning: regression
  - least-squares
    - line fitting [video]
    - line fitting, y = mx [video]
    - line fitting, y = mx + b [video]
    - parabola fitting, y = ax² + bx + c [video]
  - weighted least-squares
    - line fitting [video]
    - line fitting, y = mx + b [video]
    - line fitting, implementation [video]
  - total least-squares
    - line fitting [video]
    - line fitting, ax + by = 0 [video]
    - line fitting, implementation [video]
  - least-squares, summary [video]
  - iterative least-squares
    - intuition [video]
    - quadratic form [video]
    - steepest descent [video]
    - steepest descent, implementation [video]
    - conjugate gradient descent [video]
    - gradient descent [video]
- supervised learning: classification
  - linear classifiers
    - least-squares classifiers [video]
    - logistic regression [video]
    - linear discriminant analysis [video]
    - receiver operating curve (ROC) [video]
    - linear discriminant analysis, implementation [video]
    - multiclass classifier [video]
  - support vector machines (SVM)
    - margins [video]
    - maximizing margins [video]
    - linear SVM [video]
    - slack variables [video]
    - nonlinear SVM [video]
  - artificial neural networks
    - neurons [video]
    - delta rule [video]
    - sigmoidal neurons [video]
    - xor [video]
    - hidden layers [video]
    - xor with hidden layers [video]
    - universal approximation theorem [video]
    - backpropagation [video]
    - convolutional neural networks [video]
- unsupervised learning
  - clustering
    - k-means [video]
    - k-means, implementation [video]
  - expectation/maximization (EM)
    - EM [video]
    - E-step [video]
    - M-step [video]
    - EM, implementation [video]
  - principal component analysis (PCA)
    - canonical basis [video]
    - covariance matrix [video]
    - covariance matrix eigenvectors [video]
    - PCA, implementation [video]
    - PCA, computational considerations [video]
    - PCA for face recognition (eigenfaces) [video]
  - t-distributed stochastic neighbor embedding (tSNE)
    - tSNE [video]
    - tSNE implementation [video]
Closing
- parting thoughts [video]