CATALOG DESCRIPTION: Advanced topics in computer vision including low-level vision, geometrical and 3D vision, stereo, 3D scene reconstruction, motion analysis, visual tracking, object recognition and human motion analysis, capturing and recognition, with the applications to video processing and vision-based modeling and interaction.

RECOMMENDED TEXT: D. Forsyth and J. Ponce, Computer Vision -- A Modern Approach , Prentice Hall, 2002.

REFERENCE TEXTS:

•  O.D.Faugeras, "Three Dimensional Computer Vision: A Geometric Viewpoint", MIT Press, 1993

•  R. Duda, P. Hart and D. Stork, "Pattern Classification", John Wiley&Sons, 2001.

•  B.Horn, "Robot Vision", MIT Press, 1986

•  Y. Ma, S. Soatto, J. Kosecka and S. Sastry , “An Invitation to 3-D Vision: From Images to Geometric Models”, Springer, 2004

READINGS : Papers from journals and conference proceedings.

COURSE DIRECTOR: Prof. Ying Wu

COURSE GOALS: To gain a profound understanding of the theories, algorithms of the state-of-the-art of computer vision, various mathematical approaches, and the applications to video processing and vision-based modeling and interaction. This is a research-orientated course.

PREREQUISITES BY COURSES: EECS 230, EECS 302, and EECS 332, or equivalent

PREREQUISITES BY TOPIC:

(1) Linear algebra and probability 
(2) Digital image analysis and processing 
(3) C/C++, MATLAB

DETAILED COURSE TOPICS:

(1) Camera models and image formation 
(2) Low-level visual processing and texture 
(3) Radiometry, BRDF, color and lighting 
(4) Segmentation and grouping 
(5) Optical flow and motion analysis 
(6) Visual tracking, Kalman filtering and Sequential Monte Carlo 
(7) Deformable and articulated motion analysis 
(8) Geometry, stereo, structure from motion 
(9) Scene modeling and 3D reconstruction 
(10) Useful statistical learning techniques (including HMM, Bayesian Nets, SVM, ICA) 
(11) Pattern recognition and object recognition 
(12) Application to vision-based human computer interaction (human motion, face/gesture recognition)

LAB PROJECTS:

(1) Implementation of edge and corner detectors 
(2) Implementation of color-based image segmentation 
(3) Implementation of flow computation 
(4) Implementation of image feature matching 
(5) Ball tracking by Kalman filtering 
(6) Implementation of PCA and LDA 
* MATLAB codes are acceptable.

FINAL PROJECT TOPICS:

(1) Surveying the state-of-the-art of any topic in vision 
(2) Facial feature tracking (test sequence provided) 
(3) Face detection and recognition (test sequence provided) 
(4) Multiple moving object tracking (test sequence provided) 
(5) Simple gesture/action recognition (test sequence provided) 
(6) Stereo, 3D reconstruction and image mosaic (test sequence provided) 
(7) Algorithms for some learning tasks (data sets provided) 
(8) Image or texture classification (databases provided) 
(9) Implementation of any interesting vision algorithm 
(10) Anything that you think is NEW!

GRADES:

(1) Homeworks and labs --- 30% 
(2) Final Projects and presentations --- 70%

COURSE OBJECTIVES: When a student completes this course, s/he should be able to: 
(1) Understand the core theories and algorithms of computer vision and video processing

(2) Understand the state-of-the-art of computer vision and image/video processing, 
(3) Perform image feature detection/matching and 3D scene reconstruction, 
(4) Perform visual tracking and motion analysis, 
(5) Perform visual object recognition and motion recognition tasks, 
(6) Understand the applications such as vision-based modeling and interaction.