University Homepage School Homepage School Contacts School Search

Advanced Vision - Syllabus / Lecture Plan

1. Imaging Overview / 2D Geometry review / 2D geometric modeling
     Lecture Note 0: [PDF]  Time lapse video example: [WMV]  Lecture Problem Solution: [PDF]
     Reading:
   • Ensure familarity with the IVR course vision techniques
   • Course textbooks (one of):
     • Davies: 1.1-1.3 / 2.1 / 21.2
     • Morris: 1.1-1.3 / 2.1-2.3 / 2.5 / 3.6.1
     • Forsyth/Ponce: 1.3-1.4 / 2.1
     • Szeliski: 2
   
     Examples (in matlab): RGB Image display; RGB to Grayscale conversion
     Online: Color images; Greyscale images; Pixels; Images; Informatics New Building Timelapse Videos (updated daily)
   
   
2. System 1 - Geometric Recognition of Flat Parts (2D):
     Lecture Note 1: [PDF]
     Model-based object recognition: Boundary Extraction, Boundary Segmentation,
     Model Matching, Pose Estimation, Verification
     Reading:
   • Course textbooks (one of):
     • Davies: 4.3-4.4 / 6.11 / 7.2 / 22.1-22.4 / 22.7-22.8 / 22.12 (Chapter 5 - edge detection)
     • Morris: 4.1-4.2 / 5.2 / 5.4 / 6.3-6.3.2
     • Forsyth/Ponce: Chapter 8 - edge detection
     • Szeliski: 4.2, 6.1, 6.2, 14.3
   
     Examples (in matlab): mybwperim.m; Example code for lecture content;
     Online: Thresholding; Edge Grouping Introduction; Boundary Connectivity; Edge Tracking;
        Boundary representation introduction; Parametric boundary representations;     Tree Search Methods for Model Matching;
       Interpretation Tree Algorithm (N.B. example is best first search, in lectures we did depth first search);
       Online java applet example of Interpretation Tree based edge matching (explanation of task);
   
   
3. System 2 - Tracking a falling 2D object : Condensation Tracking
     Lecture Note 2: [PDF] Ball Condensation Tracking [MPG] Adaptive Change Detection [MPG]
     Reading:
   • Course textbooks (one of):
     • Davies: 18.13 (18.8-18.9 for background)
     • Morris: 9.3.1 / 9.3.2 / 9.4.3 / 9.4.4
     • Forsyth/Ponce: 14.3 / 17.3 / 17.5 (Chapter 17 intro, 17.1-17.2 for background) / Section 2.2 of this extra chapter (not in book)
     • Szeliski: -
   
     Examples (in matlab): Example code for ball tracking Online Demos
     Online: Kalman Filtering - concept  detail
        Traffic and Pedestrian Tracking; People Tracking [MPG] [MPG]
     Online:    The Condensation Algorithm;
   
   
4. System 3 - Persistent Tracking and Behaviour Recognition
     Lecture Note 3: [PDF]
     Reading:
   • A. A. Efros, A. C. Berg, G. Mori, J. Malik, "Recognizing Action at a Distance", Proc. IEEE Int. Conf. on Computer Vision, Nice, Vol 2, pp 726-733, Oct 2003.
   • Pedro M. Jorge, Arnaldo J. Abrantes, Jorge S. Marques. "On-line Object Tracking with Bayesian Networks". International Workshop on Image Analysis for Multimedia Interactive Systems, Lisbon, April 2004.
   
   
   
5. System 4 - Modeling and Recognizing Classes of Shapes : PCA and PDM
     Lecture Note 4: [PDF]
     Reading:
   • Course textbooks (one of):
     • Davies: 24.10 (PCA) 24.12 (Face Recongition)
     • Morris: 3.4.4 (PCA) 7.3.1-7.3.2 (PCA images)
     • Forsyth/Ponce: 22.3.1-22.3.3
     • Szeliski: 14.2
   
     Examples (in matlab): Example code for "T" recognition via PCA/PDM; Example code for Eigen faces (external)
     Online:      PCA; Point Distribution Models (PDM); PCA Representations;
        SVD (general); Gaussian Noise
        Eigen faces : 1 2 3
        Eigen faces : original Turk/Pentland paper (1991)
     Online: Mahalanobis Distance;
   
   
6. System 5 - 3D Object Recognition from Stereo Image Data : Feature Detection
     Lecture Note 5: [PDF]
     Reading:
   • Course textbooks (one of):
     • Davies: 3.2 / Chapter 5 / 7.1.1 / 16.3.1 / 16.3.2 / 21.3-21.4 / 21.7 / 21.12-21.13 / Appendix A.6
     • Morris: 3.3.3 / 3.5.2 / 5.5.3 / 6.5.2
     • Forsyth/Ponce: 3.2 / 7.1.1/ 8.2.3 / 10.1.1 / 11.1-11.4 / 15.5.2
     • Szeliski: 4,6.1,6.2,11
   
     Examples (in matlab): edge function; smoothing (via fspecial function);   Example code for stereo wedge recognition
     Online: Edge Detectors; Canny Edge Detector; Gaussian smoothing;
      RANSAC; Stereo Vision Overview; Convolution - IVR lecture 1 Stereo Vision Overview; Stereo Imaging Intro.;
        Another view on Stereo; Ballard & Brown - Section 3.4.1 ; Dense Stereo Matching; Online Comparison of Dense Stereo Approaches;
        Dimensional Imaging Website, SIFT
   
   
7. System 6 - 3D Object Recognition from Range Data : Range Sensors & Differiential Geometry
     Lecture Note 6: [PDF] 3D cola bottle [MPG] Laser Scanner [WMV]
     Reading:
   • Course textbooks (one of):
     • Davies: 16.1-16.2 16.8, 16.9 (16.10-16.12 - background only) 16.13 16.14
     • Morris: nothing appropriate
     • Forsyth/Ponce: 21.1 / 21.2 / 21.4.1
     • Szeliski: 12.2
   
     Examples (in matlab): Example code for wedge recognition;
         Movie of matlab code execution and segmentation by region growing [MPG
     Online:     Range Images (depth maps) 1;     Range Images 2;
       Time of Flight Scanners;     Triangulation based scanners;
       Section 8.7 of Machine Vision - David Vernon;
       Curvature Classification; Planar Fitting interpetation tree + lecture 3 links; Scalar Triple Product
       [Breckon/Fisher '04] - Detecting changes in 3D scenes using simple methods
   
   

There are also some review materials based on the IVR course, useful for reviewing some MATLAB and elementary image analysis:

1. Image capture and flat part recognition   [PDF]
   
2. Thresholding and background removal   [PDF]
   
3. Invariant Shape Descriptors   [PDF]
   
4. Basic Object Recognition   [PDF]
   
5. Active Vision Techniques   [PDF]
   
6. Visual Servoing   [PDF]
   

[AV home page]

---

This page is maintained by the course lecturer, Bob Fisher, rbf@inf.ed.ac.uk, room IF 1.26, ext 513441