Computer Vision 2
Semester: |
SS 2014 |
Type: |
Lecture |
Lecturer: |
|
Credits: |
Find a list of current courses on the Teaching page.
Type |
Date |
Room |
|---|---|---|
| Lecture | Mo, 14:15 - 15:45 | UMIC 025 |
| Lecture/Exercise | Do, 14:15 - 15:45 | UMIC 025 |
Description
The lecture will cover advanced topics in computer vision. A particular focus will be on state-of-the-art techniques for object detection, tracking, body pose and activity analysis. There will be regular exercises accompanying the lecture.
Further details will be announced in the first lecture on Tuesday, 15. April.
Literature
In the last decades, Computer Vision has evolved into a rapidly growing field with research going into so many directions that no single book can cover them all. Some basic material can be found in the following book:
- Computer Vision - A Modern Approach, D. Forsyth, J. Ponce, Prentice Hall, 2002
- An Invitation to 3D Vision, Y. Ma, S. Soatto, J. Kosecka, S. Sastry, Springer, 2003
However, a good part of the material presented in this class is the result of very recent research, so it hasn't found its way into textbooks yet. Wherever research papers are necessary for a deeper understanding, we will make them available on this web page.
Matlab Resources
- Matlab Online Reference Documentation
- Getting started with Matlab
- Techniques for improving performance
- A useful Matlab Quick-reference card (in German).
| Date | Title | Content | Material |
|---|---|---|---|
| Introduction | What is Tracking? | ||
| Exercise 0 | Intro Matlab | ||
| Background Modeling | MoG Background Model, Online Adaptation, Non-parametric Models | ||
| Template based Tracking | LK Tracking, fast template matching, Affine LK, Line Tracking, Model based Tracking | ||
| Color based Tracking | Mean-Shift Tracking, CAMshift, Comaniciu's Kernel-based Object Tracking | ||
| no class | Workers' Day | ||
| Contour based Tracking | Deformable Contours, Energy formulation, Greedy approach, Dynamic Programming approach, Level Sets | ||
| Tracking by Online Classification | Tracking as Online Classification problem, Online Boosting, Online Feature Selection, Drift, Semi-Supervised Boosting, TLD | ||
| Exercise 1 | Background Modeling, Mean-Shift Tracking | ||
| Tracking by Detection | Tracking-by-Detection, State-of-the-Art Detectors: HOG, DPM, Viola-Jones, Integral Channel Features, VeryFast, Roerei, Hough Forests | ||
| Bayesian Filtering I | Tracking with Linear Dynamic Models, Tracking as Inference, Prediction/Correction, Kalman Filter | ||
| Bayesian Filtering II | Extended Kalman Filter, Particle Filter, Case studies | ||
| Bayesian Filtering III | Particle Filter Details, Sequential Importance Sampling, Reweighting, Proposal Distributions | ||
| no class | (Ascension) | ||
| no class | (DIES RWTH Sports Day) | ||
| Exercise 2 | Bayesian Filtering | ||
| no class | (Excursion week) | ||
| no class | (Excursion week) | ||
| Multi-Object Tracking I | Introduction, Data Association Ambiguities, Gating, NN Filter, Track Splitting Filter | ||
| no class | (Corpus Christi) | ||
| no class | (CANCELLED!) | ||
| no class | (CANCELLED!) | ||
| Multi-Object Tracking II | Multi-Hypothesis Tracking (MHT) | ||
| Multi-Object Tracking III | Tracking as Linear Assignment Problem, Min-Cost Network Flow Optimization | ||
| Articulated Tracking I | Body Pose Estimation as High-Dimensional Regression, Synthetic Training, Latent Variable Models, Gaussian Process Regression | ||
| Articulated Tracking II | Pictorial Structures, Kinematic Tree Prior, Likelihood Models, Max-Sum Algorithm, Efficient Inference | ||
| Repetition | - | ||
| Exercise 3 | Multi-Object Tracking, Articulated Tracking |
