This course provides the basics behind computational approaches to working with images, starting with image acquisition (i.e. cameras/sensors), proceeding through understanding different representations (bitmap/raster, spectral), image manipulations (in-filling, mosaicing, etc.).
Please refer to the course information sheet for details of the course.
(The schedule below is preliminary, and subject to change)
|1 (Jan. 9)||Introduction; Cameras and Images: Understanding digital images; basic camera controls; color image acquisition; image noise||Sections 1.1-1.2, 2.1, 2.2, 2.4.2 (only paragraph entitled "silicon sensors"), 2.6.2 from Castleman book||clarkvision.com: A very comprehensive website about photography, cameras and how to characterize their properties|
|2 (Jan. 16)||HDR Imaging and Alpha Matting: Computing camera response functions from images; the matting equation||Sections 1 and 2, up to Eq (2), from the Debevec 1997 Siggraph paper in Resources/Readings. As you read the paper, note that film response curve and camera response curve in the case of digital cameras, are one and the same.||The HDRShop home page
Rendering with Natural Light (a movie that uses high-dynamic-range photography to capture outtdoor illumination and re-use it for image synthesis)
|3 (Jan. 23)||Computing 1D image derivatives: Least-squares polynomial fitting, intensity derivatives, weighted least squares, RANSAC||To run the demo shown in class: (1) unpack the file polydemo.zip in the Demo Code directory, (2) run MATLAB, (3) change the current MATLAB directory to the directory you unpacked the code, (4) type polydemo at the matlab prompt. You should run the demo for a variety of fits (LS, WLS, 1st degree, 2nd degree, etc) to see their effect.|
|4 (Jan. 30)||Edge detection: Local analysis of 1D and 2D image patches: the Image Gradient; case study: Painterly Rendering||See Litwinowicz paper on painterly rendering in Resources/Readings (this is not required reading).|
|5 (Feb. 6)||Corner detection, Intelligent Scissors, Seam Carving: Relation between local shape near extrema and the eigenvectors/eigenvalues of the Hessian; relation between eigenvalues & the trace & determinant of a matrix; localizing edges as zero crossings of the Laplacian; the Lowe feature detector: finding non-cylindrical points through eigenvalue analysis of the Hessian of the Laplacian||See Mortensen paper on intelligent scissors in Resources/Readings (this is not required reading).
This paper and the technique will be covered in one of the forthcoming tutorials.
|6 (Feb. 13)||Template matching, correlation and patch-based image processing: Representing images as vectors; evaluating similarity using RMS distance error, cross-correlation and normalized cross-correlation; non-local means denoising||To run the demo shown in class: (1) unpack the file corrdemo.zip in the Demo Code directory, (2) run MATLAB, (3) change the current MATLAB directory to the directory you unpacked the code, (4) type corrdemo at the matlab prompt.|
|7 (Feb. 27)||Principal Component Analysis||Section 13.6 from Castleman||To run the demo shown in class: (1) unpack the file recognition_demo.zip in the Demo Code directory, (2) run MATLAB, (3) change the current MATLAB directory to the directory you unpacked the code, (4) type pca_recdemo at the matlab prompt.|
|8 (Mar. 6)||Gaussian Pyramids||Original paper by Burt and Adelson on the Gauss/Laplacian pyramids in Sources/Readings. You should read up to, but not including, section entitled Entropy.||The Matlab demo scale_space1D_demo.zip in the Demo Code directory on Dropbox shows how a 1D image changes as we smooth it with a sequence of Gaussians of increasing standard deviation.|
|9 (Mar 13)||The Haar Wavelet Transform: Wavelet compression of 1D and 2D images||The paper on the Haar Wavelets by Stollnitz et al in Resources/Readings.||The Matlab demo wavedemo.zip in the Demo Code/ directory is the demo shown in class. Type wavedemo at the Matlab prompt to run it.|
|10 (Mar 20)||Polynomial fitting vs. correlation; Matching images using SIFT: Analysis of WLS polynomial fitting and image smoothing as a template matching operation; template matching expressed as a multiplication of an image with a Toeplitz matrix; Gaussian image smoothing; SIFT-based feature detection; the SIFT descriptor; image matching using SIFT||Sections 1-3 of the Lowe paper on SIFT found in the Readings/ directory on Dropbox.||Web page on SIFT (with demo code)|
|11 (Mar. 27)||Homogeneous coordinates: Homography-based image warping|
|12 (Apr. 3)||Learning Image Filters with Neural Networks: Convolutional Neural Networks|
|LEC101||Wednesday 14:00 — 16:00, GB 119||Wednesday 16:00 — 17:00 , BA 2283||Yani Ioannou|
|LEC2501 / LEC5101||Wednesday 18:00 — 20:00, BA 1190||Monday 18:00 — 19:00, BA 2283||Yawen Ma|
|Both office hours are open to students from any section.|
The readings are all available here