Image Understanding with Organic Computing Rolf P. W urtz - - PowerPoint PPT Presentation

▶
image understanding with organic computing
SMART_READER_LITE
LIVE PREVIEW

Image Understanding with Organic Computing Rolf P. W urtz - - PowerPoint PPT Presentation

Feb 05, 2024 576 likes •839 views

Neuroinformatik / Ruhr-Universit at Bochum Rolf P . W urtz Image Understanding with Organic Computing Rolf P. W urtz Ruhr-Universit at Bochum Institut f ur Neuroinformatik http://www.neuroinformatik.rub.de

slide-1
SLIDE 1

Neuroinformatik / Ruhr-Universit¨ at Bochum

Rolf P . W¨ urtz

Hannover, 2010-02-23

Image Understanding with Organic Computing

Rolf P. W¨ urtz

Ruhr-Universit¨ at Bochum Institut f¨ ur Neuroinformatik http://www.neuroinformatik.rub.de rolf.wuertz@neuroinformatik.rub.de rolf.wuertz@organic-computing.org

slide-2
SLIDE 2

Overview

Neuroinformatik / Ruhr-Universit¨ at Bochum

Rolf P . W¨ urtz

Image Understanding with Organic Computing Hannover, 2010-02-23

  • Introduction
  • Problem of image understanding
  • Controlled generalization in face recognition
  • General object recognition
  • Learning of articulated models
  • Where to go from here
slide-3
SLIDE 3

Complexity problems

Neuroinformatik / Ruhr-Universit¨ at Bochum

Rolf P . W¨ urtz

Image Understanding with Organic Computing Hannover, 2010-02-23

  • Artificial systems are rapidly getting too complex to understand
  • Desirable are complex systems with trivial interfaces
  • This requires restrictions on possible behaviors
  • This asks for self-organization
slide-4
SLIDE 4

Image understanding . . .

Neuroinformatik / Ruhr-Universit¨ at Bochum

Rolf P . W¨ urtz

Image Understanding with Organic Computing Hannover, 2010-02-23

. . . means establishing a symbolic description.

slide-5
SLIDE 5

Image understanding . . .

Neuroinformatik / Ruhr-Universit¨ at Bochum

Rolf P . W¨ urtz

Image Understanding with Organic Computing Hannover, 2010-02-23

. . . in the brain is done under additional assumptions.

slide-6
SLIDE 6

Image understanding . . .

Neuroinformatik / Ruhr-Universit¨ at Bochum

Rolf P . W¨ urtz

Image Understanding with Organic Computing Hannover, 2010-02-23

. . . requires extensive world knowledge.

slide-7
SLIDE 7

Image understanding . . .

Neuroinformatik / Ruhr-Universit¨ at Bochum

Rolf P . W¨ urtz

Image Understanding with Organic Computing Hannover, 2010-02-23

. . . has a local-global problem.

slide-8
SLIDE 8

Face recognition

Neuroinformatik / Ruhr-Universit¨ at Bochum

Rolf P . W¨ urtz

Image Understanding with Organic Computing Hannover, 2010-02-23

= =

Different situations yield very different images.

slide-9
SLIDE 9

Invariance problem

Neuroinformatik / Ruhr-Universit¨ at Bochum

Rolf P . W¨ urtz

Image Understanding with Organic Computing Hannover, 2010-02-23

All the same?

slide-10
SLIDE 10

Invariance . . .

Neuroinformatik / Ruhr-Universit¨ at Bochum

Rolf P . W¨ urtz

Image Understanding with Organic Computing Hannover, 2010-02-23

6 = 6 . . . is task-dependent.

slide-11
SLIDE 11

List of tasks

Neuroinformatik / Ruhr-Universit¨ at Bochum

Rolf P . W¨ urtz

Image Understanding with Organic Computing Hannover, 2010-02-23

  • We do need a formal model of images, but we don’t have any
  • We do not even have a formalization of the problem
  • Required is an imitation of human capability
  • Identify constraints of visual data autonomously
  • Learn computer vision routines from examples
  • Start a positive feedback loop of learning vision
  • Control generalization
slide-12
SLIDE 12

Controlled generalization

Neuroinformatik / Ruhr-Universit¨ at Bochum

Rolf P . W¨ urtz

Image Understanding with Organic Computing Hannover, 2010-02-23

  • Neural networks learn complicated functions from examples
  • They can generalize, but not always in the desired way
  • Visual invariances must be built in explicitly

(Neocognitron, Convolutional NN, . . . )

  • Exception: Slow feature analysis (Wiskott & Sejnowski, 2002)
  • Goal: Learn generalization dimensions from examples!
slide-13
SLIDE 13

Bunch Graphs

Neuroinformatik / Ruhr-Universit¨ at Bochum

Rolf P . W¨ urtz

Image Understanding with Organic Computing Hannover, 2010-02-23

general face

bunch graph Gabor wavelet jet bunch of jets a: b: c:

Wiskott et al., 1997

slide-14
SLIDE 14

Graph similarity

Neuroinformatik / Ruhr-Universit¨ at Bochum

Rolf P . W¨ urtz

Image Understanding with Organic Computing Hannover, 2010-02-23

  • Jet similarity function SJ
  • Probe graph P with N nodes Pn
  • Gallery graphs Gg with N nodes Gg,n each

grec = arg max

g

1 N

N

  • n=1

SJ(Pn, Gg,n) .

slide-15
SLIDE 15

Face Graphs

Neuroinformatik / Ruhr-Universit¨ at Bochum

Rolf P . W¨ urtz

Image Understanding with Organic Computing Hannover, 2010-02-23

slide-16
SLIDE 16

Pose Variation

Neuroinformatik / Ruhr-Universit¨ at Bochum

Rolf P . W¨ urtz

Image Understanding with Organic Computing Hannover, 2010-02-23

PM+45 PM+00 PM−45

slide-17
SLIDE 17

Pose Variation

Neuroinformatik / Ruhr-Universit¨ at Bochum

Rolf P . W¨ urtz

Image Understanding with Organic Computing Hannover, 2010-02-23

Probe Gallery Similarity? Gg P . . . . . .

slide-18
SLIDE 18

Rank Correlation

Neuroinformatik / Ruhr-Universit¨ at Bochum

Rolf P . W¨ urtz

Image Understanding with Organic Computing Hannover, 2010-02-23

Model Probe Gallery SP SG π = [7, 3, 9, . . .] γ = [7, 9, 3, . . .] γ = [2, 1, 35, . . .] . . . . . . . . . . . .

slide-19
SLIDE 19

Neural rank list similarity

Neuroinformatik / Ruhr-Universit¨ at Bochum

Rolf P . W¨ urtz

Image Understanding with Organic Computing Hannover, 2010-02-23

aj wj E I E =

K

  • j=1

exp

  • −order(aj)

λ

  • wj

wj = 1 K exp

  • −order(bj)

λ

  • Thorpe et al., 2001
slide-20
SLIDE 20

Neural rank list similarity

Neuroinformatik / Ruhr-Universit¨ at Bochum

Rolf P . W¨ urtz

Image Understanding with Organic Computing Hannover, 2010-02-23

π(m) wm,g Ag wm,g = 1 NM exp

  • −γg(m)

λ

  • Ag =
  • m

exp

  • −π(m)

λ

  • wm,g

= 1 NM

  • m

exp

  • −π(m) + γg(m)

λ

  • = Sneural(γg, π)

grec = arg max

g

1 N

  • n

Sneural(γg,n, πn)

M¨ uller and W¨ urtz, ICANN 2009

slide-21
SLIDE 21

Face recognition Mainz Hbf.

Neuroinformatik / Ruhr-Universit¨ at Bochum

Rolf P . W¨ urtz

Image Understanding with Organic Computing Hannover, 2010-02-23

slide-22
SLIDE 22

CAS-PEAL Database

Neuroinformatik / Ruhr-Universit¨ at Bochum

Rolf P . W¨ urtz

Image Understanding with Organic Computing Hannover, 2010-02-23

PM+45 FM+00 FM−45 FM−90 PM+00 FD+00 FD−45 FD−90 PM−45 FU+00 FU−45 FU−90

slide-23
SLIDE 23

Rank Correlation: Results

Neuroinformatik / Ruhr-Universit¨ at Bochum

Rolf P . W¨ urtz

Image Understanding with Organic Computing Hannover, 2010-02-23

Pose Illumination Recognition percentage with given situation 99.02 89.01 Percentage of correct situation estimation 99.89 ± 0.09 91.96 ± 0.89 Recognition percentage with automatically determined situation 97.75 ± 0.50 89.97 ± 1.36 Best recognition percentage reported in database description 71 51

slide-24
SLIDE 24

Rank Correlation: Early stopping

Neuroinformatik / Ruhr-Universit¨ at Bochum

Rolf P . W¨ urtz

Image Understanding with Organic Computing Hannover, 2010-02-23

0.2 0.4 0.6 0.8 1 20 40 60 80 100

Recognition rate Spike number pose illumination

slide-25
SLIDE 25

Thanks to

Neuroinformatik / Ruhr-Universit¨ at Bochum

Rolf P . W¨ urtz

Image Understanding with Organic Computing Hannover, 2010-02-23

Marco M¨ uller Rank correlation memory G¨ unter Westphal Object recognition Thomas Walther Body tracking Manuel G¨ unter Statistical face recognition Markus Lessmann Scene analysis Oliver Lomp Neuronal dynamics Mathis Richter Clustering of image patches Guillermo Donatti Object memory, Neural Map DFG, EU, NRW Funding All of you Attention