Date Topic Readings and Annoucements
1/30 Introduction and logistics

Module 1: Image Formation
2/4 Light and color: I
- Spectral basis of light
- Color perception in the human eye

2/6 Class cancelled, snow day

2/11 Light and color: II
- Tristimulus theory and color spaces
- Color phenomenon

2/13 Image formation: I
- Pinhole camera model
- Qualitative properties
- Cameras with lens
- Lens phenomenon

2/18 Image formation: II
- Lens flaws and modern lenses
- Analog and digital color sensing
- Alignment and demosacing

2/20 No Class, Monday schedule

Module 2: Image Processing
2/25 Python tutorial

2/27 Image processing: I
- Spatial and bightness quantization
- Improving brightness and contrast
- The convolution operation
- Denoising

3/4 Image processing: II
- Sharpening
- Hybrid Images
- Edge detection
  • Homework 3 released (due 3/27)
  • Szeliski book, Chapter 3

3/6 Local features: I
- Role of local features
- Harris corner detector

3/11 Local features: II
- Harris corner detector
- Invariance and equivariance

3/13 Midterm in class

3/18 Spring break

3/20 Spring break

3/25 Local features: III
- Invariance and equivariance
- Blob detector
- Eliminating rotational ambiguity
- The SIFT descriptor

3/27 Matching and alignment: I
- The SIFT descriptor
- Matching and ratio test
- Model fitting using RANSAC

4/1 Matching and alignment: II
- Image transformations
- Estimating transformations from matching
- Image warping
- Modern approaches
  • Szeliski book, Chapter 8

4/3 Optical flow
- Motion field and optical flow
- Lucas-Kanade optical flow
- Alternate flow techniques
- Depth estimation, point tracking, video interpolation

Module 3: Image Understanding
4/8 Recognition by alignment
- Instance matching
- Vocabulary trees
- Beyond instance matching

4/10 Visual recognition
- What is recognition?
- Brief history of recognition
- Current trends
  • Slides
  • Szeliski book, Chapter 6

4/15 Classical machine learning: I
- The machine learning framework
- Nearest neighbor classifiers
- Hyperparameters search

4/17 Classical machine learning: II
- Linear classifiers
- Loss function, regularization, optimization

4/22 Image representations
- Role of representations
- Classical representations (HOG and BoVW)

4/24 Deep learning: I
- Multi-layer perceptrons
- Activations, Loss functions, optimization
- Practical issues
- Convolutional networks

4/29 Deep Learning: II
- Convolutional networks
- LeNet
- AlexNet and ImageNet challenge

5/1 Deep Learning: III
- AlexNet and ImageNet challenge
- Visualizing deep networks
- Transfer learning
- Modern CNNs

5/6 Object detection
- Sliding-window detectors
- Region-based detectors
- Datasets and benchmarks

5/8 Guest lecture by Aaron Sun

5/14 Final, 1- 3pm, LGRC A301