Statistical and Computational Models in Vision

This is a graduate level course on computer image analysis. Upon completion, students should be able to sketch a global picture of the computer vision field and have learned fundamental concepts and advanced methods in statistical modeling of images, stochastic computation, and projective geometry.

Instructor: Pedro M. Q. Aguiar, ISR-IST, contact: aguiar@isr.ist.utl.pt

Reading material: There is no textbook for the course. The readings are research papers and selected chapters/sections from various books. The majority of these papers are either linked through the course web page or provided as handouts.

Pre-requisites: Background in computer vision, pattern recognition, statistics. Previous experience on computer vision/image processing projects. MATLAB programming skills.

Grading: Students will be graded according to i) lecture attendance and discussion and ii) a presentation on a specific topic. i) In order to participate discussions effectively, students should read the materials before attending the lectures. ii) Students should identify a topic of their interest as soon as possible and prepare a 60 min. presentation on that topic.

Tentative syllabus:

Introduction. What is vision? Scientific approaches to vision. State of the art. Open problems. Vision as Bayesian inference.

"A Bayesian formulation of visual perception", D.C. Knill, D. Kersten, and A. Yuille, chapter in "Perception as Bayesian Inference", Cambridge University Press, 1996, (handout).

Image modeling and representation. Statistical models.

"What is the goal of sensory coding?", D.J. Field, Neural Computation, 6:559-601, 1994.
"Sparse coding with an over-complete basis set: A strategy employed by V1?", B.A. Olshausen and D.J. Field, Vision Research, 37:3311-3325, 1997, [.ps].
"Minimax entropy principle and its applications to texture modeling", S.C. Zhu, Y.N. Wu, and D.B. Mumford, Neural Computation, 1997, [.ps].
"Edge co-occurence in natural images predicts contour grouping performance", W.S. Geisler, J.S. Perry, B.J. Super, D.P. Gallogly, Vision Research, 41:711-724, 2001.
"Statistical modeling and conceptualization of visual patterns", S.C. Zhu, Submitted to PAMI, 2002, [.ps].

Stochastic computation. Diffusions, jumps, MCMC.

"Mean shift, mode Seeking, and clustering", Y. Cheng, PAMI, 17(8):790-799, 1995.
"Region competition: unifying snakes, region growing, and Bayes/MDL for multi-band image segmentation", S.C. Zhu and A.L. Yuille, PAMI, 1996, [.ps].
"Level set methods: an act of violence", J.A. Sethian, A tutorial on level set methods, 1997, [.ps], [web link].
"Image segmentation by data-driven Markov Chain Monte Carlo", Z.W. Tu and S.C. Zhu, PAMI, 2002, ICCV 2001, [.ps].

Scene geometry. Projective geometry, stereo, multiview geometry, SFM.

"Shape and motion from image streams: a factorization method", C. Tomasi and T. Kanade, IJCV, 1992, (handout).
"Self-Calibration and Metric Reconstruction in spite of Varying and Unknown Internal Camera Parameters", Pollefeys, Koch and Gool, IJCV, 32(1), 1999, [pdf].
"The Manhattan World assumption: regularities in scene statistics which enable Bayesian inference", J. Coughlan and A. Yuille, NIPS 2000, [.ps], ICCV 1999, [.pdf], [web link].
"Structure from motion without correspondences", F. Dellaert S. M. Seitz C. E. Thorpe S. Thrun, CVPR 2000, [.pdf].
"The Space of All Stereo Images", S. Seitz, J. Kim, IJCV, 2001, [pdf].

Motion analysis and tracking. Condensation, particle filtering.

"Representing moving images with layers", J. Wang and E.H. Adelson, IEEE Trans. on Image Processing, 3(5):625-638, 1994, (handout).
"Estimating optical flow in segmented images using variable-order parametric models with local deformations", M.J. Black and A. Jepson, PAMI, 1996, [pdf].
"Condensation - conditional density propagation for visual tracking", M. Isard and A. Blake, IJCV, 1998, [.ps], ECCV 1998, [.ps], [web link].

Vision with graphics and arts. Image-based rendering, non-photorealistic rendering.

"Image-based rendering", D. Forsyth and J. Ponce, chapter in "Computer Vision", [pdf].
"Plenoptic sampling", J.X. Chai, X. Tong, S.C. Chan, and H. Shum, Siggraph 2000, [pdf].
"Genealized plenoptic sampling", C. Zhang and T. Chen, CMU TR AMP01-06, 2001, [pdf].
"Non-Photorealistic Rendering", tutorial at Siggraph 19990, [pdf], [web link].