Informatik an der TU Graz - Institute for Computer Graphics and Vision

icg.tugraz.at

Informatik an der TU Graz - Institute for Computer Graphics and Vision

Bildverarbeitung und Mustererkennung

Image Processing and Pattern Recognition

710.080 2VO

710.081 1KU

Thomas Pock & Andreas Wendel Image Processing and Pattern Recognition, WS2012/2013

1


About the Teachers

Thomas Pock Andreas Wendel

Institute for Computer Graphics and Vision

• Inffeldgasse 16 / 2nd floor

• Homepage:

http://www.icg.tugraz.at/Members/{pock,wendel}

• Email: {pock,wendel}@icg.tugraz.at

• Office hours by agreement (email)

Thomas Pock & Andreas Wendel Image Processing and Pattern Recognition, WS2012/2013

2


Lecture (VO)

• Content: Advanced techniques in image processing,

computer vision and pattern recognition

• Every Monday: 14:15h-15:45h, Lecture hall i2

• Registration: TUG-Online

• Examination:

– Written, at the end of the winter term

– Oral, anytime (by agreement)

• Newsgroup: tu-graz.lv.bvme (also for exercise)

Thomas Pock & Andreas Wendel Image Processing and Pattern Recognition, WS2012/2013

3


Planned Content for 2012/2013

08.10.2012 TP&AW Introduction, Linear Filtering

15.10.2012 TP Advanced Filtering

22.10.2012 AW Edges, Structure Tensor, Hessian

29.10.2012 AW Feature Detectors

05.11.2012 AW Feature Descriptors

12.11.2012 TP Supervised classification, SVM, logistic regression

19.11.2012 AW Unsupervised Classification, K-means, GMM

26.11.2012 AW Segmentation, Hough Transform, Shapes

03.12.2012 TP Mean Shift Analysis, Wavelets

10.12.2012 TP Markov / Conditional Random Fields

17.12.2012 TP&AW Project Presentations

07.01.2013 TP PDE Methods

14.01.2013 TP Variational Methods

21.01.2013 TP&AW Project Presentations (plus 1 additional date)

Thomas Pock & Andreas Wendel Image Processing and Pattern Recognition, WS2012/2013

4


Exercise (KU)

• Practical implementations of image processing methods

• Tutors: Kerstin Hammernik, Gernot Riegler

• Email: {hammernik, riegler}@icg.tugraz.at

• Tutorial: Monday, 16:00h-17:00h, EDV Lehrsaal 1

• Software: Matlab, C++ (OpenCV)

• Hardware: Laptop, PCs, EDV Lehrsaal 1

• 4 assignments, 1 project

Thomas Pock & Andreas Wendel Image Processing and Pattern Recognition, WS2012/2013

5


Assignments

Assignment 1 08.10.2012 Filtering / Preprocessing 15%

Assignment 2 29.10.2012 Features 15%

Assignment 3 12.11.2012 Pattern Recognition 15%

Assignment 4 26.11.2012 Segmentation 15%

Project 17.12.2012 Open project 40%

• Submission:

• Document containing a detailed description of the

methodology and results

• Matlab / C++ Code

• Email to bvme@icg.tugraz.at

• Bonus points – see newsgroup for details

Thomas Pock & Andreas Wendel Image Processing and Pattern Recognition, WS2012/2013

6


Exercise (KU) - Project

• Implement your cool vision idea!

• E.g. Visual Sudoku Solver

Thomas Pock & Andreas Wendel Image Processing and Pattern Recognition, WS2012/2013

7


Exercise (KU) - Project

• Implement your cool vision idea!

• E.g. Vision-enabled garbage bin

Thomas Pock & Andreas Wendel Image Processing and Pattern Recognition, WS2012/2013

8


• Groups of 2 people

Exercise (KU) - Project

• Tutorials: is my idea feasible?

• Proposal: 1-page description by email to

bvme@icg.tugraz.at as soon as idea ready (first-come,

first-serve)

• Idea presentation (5min): Monday, 17.12.2012, 14-16h

• Progress report and support in tutorials

• Final presentation (20min): Monday, 21.01.2013, 14-18h

• Final report (just as for assignments) due 28.01.2013

• Best-Project Award!

Thomas Pock & Andreas Wendel Image Processing and Pattern Recognition, WS2012/2013

9


• Winners 2011/12:

Exercise (KU) - Project

Visual Pong by Gernot Riegler & Michael Scheer

Thomas Pock & Andreas Wendel Image Processing and Pattern Recognition, WS2012/2013

10


• Lecture:

– Slides, Matlab Files

Material

– http://www.icg.tu-graz.ac.at/courses/lv710.080/info

• Exercise:

– Assignments, Tutorials:

– http://www.icg.tu-graz.ac.at/courses/lv710.081/info

Will be updated during the semester!

Thomas Pock & Andreas Wendel Image Processing and Pattern Recognition, WS2012/2013

11


Literatur(1)

Milan Sonka, Vaclav Hlavac, Roger Boyle:

Image Processing, Analysis, and Machine Vision

(second edition) / (third edition)

PWS Publishing, 1999

ISBN 0-534-95393-X

(Lehrbuchsammlung)

http://www.icaen.uiowa.edu/~dip/LECTURE/lecture.html

Kristian Bredies, Dirk Lorenz:

Mathematische Bildverarbeitung: Einführung in

Grundlagen und moderne Theorie

Vieweg+Teubner Verlag, 2011

ISBN 3834810371

(eBook)

Thomas Pock & Andreas Wendel Image Processing and Pattern Recognition, WS2012/2013

12


Literatur(2)

Rafael C. Gonzalez, Steven L. Eddins,

Richard E. Woods:

Digital Image Processing Using MATLAB

Prentice Hall, 2003

ISBN 0130085197

(Lehrbuchsammlung)

Gary Bradski, Adrian Kaehler:

Learning OpenCV: Computer Vision with

the OpenCV Library

O’Reilly Media, 2008

ISBN 0596516134

(Lehrbuchsammlung, auch als e-Book)

Thomas Pock & Andreas Wendel Image Processing and Pattern Recognition, WS2012/2013

13


Ultimate goal: teach computers to see!

Thomas Pock & Andreas Wendel Image Processing and Pattern Recognition, WS2012/2013

14


What is so difficult about images?

• Seconds since big bang: 10 17

• Atoms in the universe: 10 80

• 65 x 65 images with 256 grey values: 10 10000

Thomas Pock & Andreas Wendel Image Processing and Pattern Recognition, WS2012/2013

15


Is the human visual system perfect?

Thomas Pock & Andreas Wendel Image Processing and Pattern Recognition, WS2012/2013

16


What is that?

Thomas Pock & Andreas Wendel Image Processing and Pattern Recognition, WS2012/2013

17


Where are the faces?

Thomas Pock & Andreas Wendel Image Processing and Pattern Recognition, WS2012/2013

18


Man or Woman?

Thomas Pock & Andreas Wendel Image Processing and Pattern Recognition, WS2012/2013

19


We can easily fool ourselves!

Thomas Pock & Andreas Wendel Image Processing and Pattern Recognition, WS2012/2013

20


How far can computers go?

Thomas Pock & Andreas Wendel Image Processing and Pattern Recognition, WS2012/2013

21


Scene Completion

Hays and Efros, 2007

Thomas Pock & Andreas Wendel Image Processing and Pattern Recognition, WS2012/2013

Hays and Efros, SIGGRAPH 2007 22


Thomas Pock & Andreas Wendel Image Processing and Pattern Recognition, WS2012/2013

Hays and Efros, SIGGRAPH 2007

23


Context Matching

Thomas Pock & Andreas Wendel Image Processing and Pattern Recognition, WS2012/2013

Hays and Efros, SIGGRAPH 2007

24


Thomas Pock & Andreas Wendel Image Processing and Pattern Recognition, WS2012/2013

Hays and Efros, SIGGRAPH 2007

25


Thomas Pock & Andreas Wendel Image Processing and Pattern Recognition, WS2012/2013

Hays and Efros, SIGGRAPH 2007

26


Thomas Pock & Andreas Wendel Image Processing and Pattern Recognition, WS2012/2013

Hays and Efros, SIGGRAPH 2007

27


Thomas Pock & Andreas Wendel Image Processing and Pattern Recognition, WS2012/2013

Hays and Efros, SIGGRAPH 2007

28


Thomas Pock & Andreas Wendel Image Processing and Pattern Recognition, WS2012/2013

Hays and Efros, SIGGRAPH 2007

29


Thomas Pock & Andreas Wendel Image Processing and Pattern Recognition, WS2012/2013

Hays and Efros, SIGGRAPH 2007

30


Dense Tracking and Mapping

Newcombe et al. 2011

Thomas Pock & Andreas Wendel Image Processing and Pattern Recognition, WS2012/2013

31


Earth viewers (3D modeling)

Image from Microsoft Bing

(see also: Google Earth)

Thomas Pock & Andreas Wendel Image Processing and Pattern Recognition, WS2012/2013

32


Photosynth

http://labs.live.com/photosynth/

Thomas Pock & Andreas Wendel Image Processing and Pattern Recognition, WS2012/2013

33


Object recognition (in mobile phones)

• This is becoming real:

– Microsoft Research

– Point & Find, Nokia

Thomas Pock & Andreas Wendel Image Processing and Pattern Recognition, WS2012/2013

34


Object recognition (in supermarkets)

LaneHawk by

EvolutionRobotics

“A smart camera is flush-mounted in

the checkout lane, continuously

watching for items. When an item is

detected and recognized, the cashier

verifies the quantity of items that

were found under the basket, and

continues to close the transaction.

The item can remain under the

basket, and with LaneHawk,you are

assured to get paid for it… “

Thomas Pock & Andreas Wendel Image Processing and Pattern Recognition, WS2012/2013

35


Optical character recognition (OCR)

Technology to convert scanned docs to text

• If you have a scanner, it probably came with OCR software

Digit recognition, AT&T labs

http://www.research.att.com/~yann/

License plate readers

http://en.wikipedia.org/wiki/Automatic_number_plate_recognition

Thomas Pock & Andreas Wendel Image Processing and Pattern Recognition, WS2012/2013

36


Face detection

• Many new digital cameras now detect faces

– Canon, Sony, Fuji, …

Thomas Pock & Andreas Wendel Image Processing and Pattern Recognition, WS2012/2013

37


Smile (?) detection

Thomas Pock & Andreas Wendel Image Processing and Pattern Recognition, WS2012/2013

38


Vision-based Biometrics

Fingerprint scanners on

many new laptops,

other devices

Face recognition systems now

beginning to appear more widely

http://www.sensiblevision.com/

Thomas Pock & Andreas Wendel Image Processing and Pattern Recognition, WS2012/2013

39


• E.g. Mobileye

Vision for Smart Cars

Vision systems currently in high-end BMW, Mercedes, Volvo

models

Thomas Pock & Andreas Wendel Image Processing and Pattern Recognition, WS2012/2013

40


Special effects: Shape capture

The Matrix movies, ESC Entertainment, XYZRGB, NRC

Thomas Pock & Andreas Wendel Image Processing and Pattern Recognition, WS2012/2013

41


Special effects: Motion capture

Pirates of the Caribbean, Industrial Light and Magic

Thomas Pock & Andreas Wendel Image Processing and Pattern Recognition, WS2012/2013

42


Special effects: Vision & Graphics

Avatar, Weta Digital

Thomas Pock & Andreas Wendel Image Processing and Pattern Recognition, WS2012/2013

43


Vision in Sports

LiberoVision

Offside detection using 3D reconstruction

Sportvision first down line

Nice explanation on

www.howstuffworks.com

Thomas Pock & Andreas Wendel Image Processing and Pattern Recognition, WS2012/2013

44


3D imaging

MRI, CT

Medical Imaging

Image guided surgery

Thomas Pock & Andreas Wendel Image Processing and Pattern Recognition, WS2012/2013

45


Vision in Space

NASA'S Mars Exploration Rover Spirit captured this westward view from atop

a low plateau where Spirit spent the closing months of 2007.

Vision systems used for several tasks:

• Panorama stitching

• 3D terrain modeling

• Obstacle detection, position tracking

• For more, read “Computer Vision on Mars” by Matthies et al.

Thomas Pock & Andreas Wendel Image Processing and Pattern Recognition, WS2012/2013

46


Micro aerial vehicles:

http://aerial.icg.tugraz.at

Robotics

Robots playing soccer:

http://www.robocup.org

Thomas Pock & Andreas Wendel Image Processing and Pattern Recognition, WS2012/2013

47


Thomas Pock & Andreas Wendel Image Processing and Pattern Recognition, WS2012/2013

48

More magazines by this user
Similar magazines