Instructors: Sumeet
Agarwal and Hiranmay Ghosh
4 credits (3-0-2)
Pre-requisites: EEL205 & EC120
Overlaps with: CSL840
I Semester 2013–14
M Th 17–18:20, II-247
How can machines or computers be equipped with the ability to process and understand the visual information in the world around us? This course will look at algorithms and models for vision. Our approach will be to think of real-world vision as a learning and inference problem from noisy data. In particular, we will focus on employing a statistical machine learning framework for the task of building classifiers for recognising various visual phenomena.
Week(s) | Topics | Slides | CVAA Chapter(s) | Other References |
1 | Introduction, History of Computer Vision | Hays' intro; Optical illusions | 1 | |
1–2 | Image formation: cameras, optics, geometry | 2 | Hartley, Minimizing Algebraic Error in Geometric Estimation Problems | |
3 | Machine learning overview: clustering, classification | Very quick introduction; Hays' ML overview; Prince Chapter 6 | — | Prince Chapter 6; Tutorial |
4–5 | Image filtering: Fourier transforms, pyramids, wavelets | 3 | Filtered image examples; Wavelet Tutorial; Viola-Jones paper | |
5–7 | Segmentation as unsupervised learning: Gaussian Mixture Models, Markov Random Fields | GMMs and EM [Hays]; MRFs and graph cuts [Hays] | 5 | GMM tutorial; MRF tutorial; GrabCut paper; Practice Sheet I |
7 | Minor project presentations | On Piazza | — | |
7–8 | Recognition: image processing, feature extraction, bag-of-words approaches | 4, 14 | Papers | |
8 | Guest lecture 1: Dr. Sunil Kopparapu | Readable Image for the Visually Impaired | — | Readable Image for the Visually Impaired |
8–9 | Feature detection: edges, lines | 4 | Papers | |
9 | Object/face/instance recognition | Recognition overview [Hays]; Local features [Hays] | 14 | SIFT paper [Lowe '99] |
10 | Context and scene understanding | Context and spatial layout [Hays] | 14 | |
10–11 | Image indexing | 4 | LSH tutorial; Similarity Search | |
11–12 | Deep learning for feature discovery and image classification | Google Brain | — | Andrew Ng's lecture notes; Google Brain paper; Deep learning tutorial |
12 | Multimedia semantics | Part I, Part II | — | Papers |
13 | Guest lecture 2: Prof. Amitabha Mukerjee | Hierarchies in representations | — | |
13 | Journal club: Aesthetics in Vision | — | Papers | |
14 | Major project presentations | On Piazza | — |