DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING

UNIVERSITY OF CALIFORNIA, SAN DIEGO

Instructor: Serge Belongie, Assistant Professor, AP&M room 4832. Office Hours: MTu 2-3pm

Class section id for CSE252: #465981.

Lecture: TuTh 5:00-6:20pm, Solis 111.

Topics to be Covered: mathematical foundations, cameras & image formation, stereopsis, motion, feature extraction, texture, image segmentation, object recognition.

Prerequisites: linear algebra, calculus, probability and statistics. Background in signal/image processing is helpful but not required.

This course makes extensive use of Matlab. Click here for information on Matlab.

Handouts/Readings:

**Assignment #0**(Introducing Matlab) [txt]- Brunelleschi and the Origin of Linear Perspective [html]
- Computing Rotations in 3D [html]
- The Interpretation of a Moving Retinal Image (
*Longuet-Higgins and Prazdny*) [pdf] - F&P Sec. 1.1, 2.1
- An Introduction to Projective Geometry (
*Birchfield*) [html] [pdf] [ps.gz] **Assignment #1**[pdf]- A Plane Measuring Device (
*Criminisi, Reid and Zisserman*) [html][ps.zip] - Perspective Transform Estimation (
*Wren*) [html] - F&P Sec. 10.1
- In Defense of the Eight Point Algorithm (
*Hartley*) [pdf] - A computer algorithm for reconstructing a scene from two projections (
*Longuet-Higgins*) [pdf] - Epipolar Geometry and the Fundamental Matrix (Ch.8 of
*Hartley & Zisserman*) [pdf] - Reconstruction from Two Calibrated Views (Ch.5 of
*Ma, Soatto, Kosecka and Sastry*) [pdf] **Assignment #2**[pdf][files]- A Fast Operator for Detection and Precise Location of Distinct Points, Corners and Centres of Circular Features (
*Förstner and Gülch*) [pdf] - Notes on Corner Detection [pdf]
- F&P Ch. 8
- Useful Matrix Formulae (
*Roweis*) [pdf] - A Computational Approach to Edge Detection (
*Canny*) [pdf] - Junction detection with automatic selection of detection scales and localization scales (
*Lindeberg*) [ps.Z] - The Design and Use of Steerable Filters (
*Freeman and Adelson*) [html] - Linear scale-space: I. Basic theory, II. Early visual operations (
*Lindeberg and ter Haar Romeny*) [html] - Deformable kernels for early vision (
*Perona*) [ps.Z][pdf] - F&P Ch. 9
- Preattentive texture discrimination with early vision mechanisms (
*Malik and Perona*) [pdf] - Non-parametric similarity measures for unsupervised texture segmentation and image retrieval (
*Puzicha et al.*) [ps.gz] - Pyramid Based Texture Analysis/Synthesis (
*Heeger & Bergen*) [pdf] - F&P Ch. 11
- Random Sample Consensus: A Paradigm for Model Fitting with Applications to Image Analysis and Automated Cartography (
*Fischler and Bolles*) [pdf] - Feature Based Methods for Structure and Motion Estimation (
*Torr and Zisserman*) [ps.gz] - An Algorithm for Associating the Features of Two Images (
*Scott and Longuet-Higgins*) [pdf] - Uncalibrated Stereo Correspondence by Singular Value Decomposition (
*Pilu*) [html] - New Algorithms for 2D and 3D Point Matching: Pose Estimation and Correspondence (
*Gold, Rangarajan, et al.*) [ps.gz] **Assignment #3**[pdf][files]- Stereo by Intra- and Inter-Scanline Search Using Dynamic Programming (
*Ohta and Kanade*) [pdf] - Depth from edge and intensity based stereo (
*Baker and Binford*) [pdf] - Computational framework for determining stereo correspondence from a set of linear spatial filters (
*Jones and Malik*) [pdf] - F&P Ch. 4
- Principles of Global Illumination (Ch.2 of
*Sillion and Puech*) [pdf] - Helmholtz Stereopsis: Exploiting Reciprocity for Surface Reconstruction (
*Zickler, Belhumeur and Kriegman*) [pdf] - An iterative image registration technique with an application to stereo vision (
*Lucas and Kanade*) [pdf] - Good Features to Track (
*Shi and Tomasi*) [ps] [pdf] Tech Report [ps] - The Principle of Least Action (Book 2, Ch.19 of
*Feynman, Leighton and Sands*) [pdf] - Optical Flow and Motion Analysis (
*Wu*) [pdf] - Motion Field and Optical Flow (Ch.12 + Appendix A.6 of
*Horn*) [pdf] - All About Direct Methods (
*Irani and Anandan*) [ps.gz] - Principal warps: thin-plate splines and the decomposition of deformations (
*Bookstein*) [pdf] - Computational vision and regularization theory (
*Poggio, Torre and Koch*) [pdf] - Regularization theory and neural networks architectures. (
*Girosi, Jones and Poggio*) [ps.gz] - Reconstructing a Smooth Surface (Sec. 4.4 of
*Hertz, Krogh and Palmer*) [pdf] - Laws of Organization in Perceptual Forms (
*Wertheimer*) [html] - F&P Ch. 14
- A Database of Human Segmented Natural Images and its Application ... (
*Martin, Fowlkes, Tal and Malik*) [pdf] [website] - Reversible Markov Chains and Random Walks on Graphs (
*Aldous and Fill*) [pdf] - Normalized cuts and image segmentation (
*Shi and Malik*) [pdf] - A Random Walks View of Spectral Segmentation (
*Meila and Shi*) [ps] - Notes on Spectral Partitioning (
*Demmel*) [html] **Assignment #4**[pdf][files]- F&P Ch. 15
- Snakes: active contour models (
*Kass, Witkin and Terzopoulos*) [pdf] - The representation and matching of pictorial structures (
*Fischler and Elschlager*) [pdf] - Deformable templates for face recognition (
*Yuille*) [pdf] - Shape Matching and Object Recognition Using Shape Contexts (
*Belongie, Malik and Puzicha*) [pdf] - Geometric Hashing: An Overview (
*Wolfson and Rigoutsos*) [pdf] - Method and Means for Recognizing Complex Patterns (
*Hough*) [pdf] - Generalizing the Hough Transform to Detect Arbitrary Shapes (
*Ballard*) [pdf] - Efficient Pose Clustering Using a Randomized Algorithm (
*Olson*) [pdf] - F&P Ch. 18
- Eigenfaces for Recognition (
*Turk and Pentland*) [pdf] - Bayesian Face Recognition (
*Moghaddam, Jebara and Pentland*) [pdf] - Recognition by linear combinations of models (
*Ullman and Basri*) [pdf] - F&P Sec. 22.3
- Shape and Motion from Image Streams: A Factorization Method (
*Tomasi and Kanade*) [pdf] - Affine Structure from Motion (
*Koenderink and van Doorn*) [pdf] - F&P Sec. 12.3, 12.4, 12.6.
**Final Exam**[pdf]

Lectures:

- April 1: Perspective Projection, Rigid Body Transformations.
- April 3: Equations of the Motion Field, Parallax, Time-to-Collision.
- April 8: Planar Transformations, 2D Homography Estimation.
- April 10: Epipolar Geometry, Essential Matrix, Fundamental Matrix.
- April 15: 3D Reconstruction from Two Calibrated Views.
- April 17: Interest Point Detection, Windowed Image Second Moment Matrix.
- April 22: Edge Detection, Quadrature Filters, Steerable Filters.
- April 24: Texture Analysis and Recognition.
- April 29: The Correspondence Problem, Sparse Stereo Matching, RANSAC.
- May 1: Dense Stereo Matching, Filter-Based Correspondence.
- May 6: Photometry, Shape from Shading.
- May 8: Measuring Optical Flow.
- May 13: Estimating Planar Transformations, Thin Plate Splines.
- May 15: Gestalt Cues for Grouping, Eigenvector-based Pairwise Clustering.
- May 20: Image Segmentation, Normalized Cuts.
- May 22: Feature-based Methods for Recognition.
- May 27: Appearance-based Methods for Recognition.
- May 29: Structure and Motion from Tracked Points.
- June 3:
*no lecture* - June 5: Review for Final Exam

Links:

- Announcements
- Perception Demos: Aperture Problem, Ternus Effect, etc.
- eigshow: interactive exploration of eigenvalues and singular values
- Matlab information
- CSE 252 e-reserves
- Camera Calibration Toolbox for Matlab
- Intel OpenCV
- CVonline
- The Computer Vision Home Page
- Related classes at UCSD: CSE 190-B, CSE 166, ECE 253a
- Slides/figures from this class, Forsyth & Ponce, Hartley & Zisserman, Pollefeys

Required textbook:

Computer Vision: A Modern Approach Forsyth & Ponce Prentice Hall ISBN no. 0130851981 |

Introductory Techniques for 3-D Computer Vision by Trucco & VerriHandy Math reference:

Multiple View Geometry in Computer Vision by Hartley & Zisserman

The Geometry of Multiple Images by Faugeras, Luong, and Papadopoulo

Vision Science: Photons to Phenomenology by Stephen E. Palmer

MathWorld

*Most recently updated on Feb. 12, 2003 by Serge Belongie.*