Lectures
You can download all of the lecture materials here. Please email me if you find any mistakes.
-
Course Introduction and 2D Fourier transform
[Intro Slides] [Analog Images: Part I]
- What is machine perception/image processing?
- The basic objects: Images
- Vector-space formulation of analog images
- 2D systems
- 2D Fourier transform
-
2D LSI Systems and Image Acquisition
[Analog Images: Part II] [Image Sampling and Acquisition: Part I]
- Orientation estimation
- Characterization of LSI systems
- Sampling theory
- Aliasing problems
-
Image Quantization and Filtering
[Image Sampling and Acquisition: Part II] [Image Filtering: Part I]
- Image quantization
- Grayscale vs. spatial resolution tradeoff
- 2D z-transforms
- Image filtering
-
Image Filtering Considerations and Binary Morphology
[Image Filtering: Part II] [Morphological Processing: Part I]
- Separability
- Boundary conditions
- Useful image processing filters
- Binary morphology
-
Practice Midterm
[Practice Midterm] [Practice Midterm Solutions]
The midterm will cover all material through the first half of week 4
-
Morphological Processing and Common Image Processing Tasks
[Morphological Processing: Part II] [Image Processing Tasks: Part I]
- Binary morphology
- Graylevel morphology
- (Local) image normalization
- Template matching and matched filters
- Edge detection and Canny’s algorithm
- What is image segmentation?
-
Image Segmentation and Directional Image Analysis/Processing
[Image Processing Tasks: Part II] [Directional Image Analysis/Processing]
- Image segmentation
- Radon transform
- Structure tensor
- Steerable filters
- Edge and ridge detection: Revisited
-
Image Reconstruction, Inverse Problems, and Least Squares
[Image Reconstruction: Part I (Slides)] [Image Reconstruction: Part I (Board)]
- Imaging as an inverse problem
- Imaging operators and modalities
- Discretization of inverse problems
- Least squares
-
Proximal Gradient Methods and Plug-and-Play for Image Reconstruction
[Image Reconstruction: Part II (Board)]
- Convexity
- Regularized least squares
- Proximal gradient methods
- Plug-and-play methods for image reconstruction