gbr07 shen gufg 01

Information about gbr07 shen gufg 01

Published on February 4, 2008

Author: Urania

Source: authorstream.com

Content

Grouping Using Factor Graph: An Approach for finding Text with a Camera Phone:  Grouping Using Factor Graph: An Approach for finding Text with a Camera Phone Huiying Shen and James Coughlan Smith-Kettlewell Eye Research Institute San Francisco, USA Slide2:  Motivation Slide3:  “Webster” Motivation More Examples:  More Examples Find street signs, restaurants, addresses Find indoor signs as landmarks and info tags Why a Cell Phone:  Why a Cell Phone Everyone has one A Smart phone is a capable computer But, it has limited resources, compared to a desktop, And, has NO float point processor A simple and efficient algorithm is needed,…… Why a Cell Phone:  Why a Cell Phone Everyone has one A Smart phone is a capable computer But, it has limited resources, compared to a desktop, And, has NO float point processor A simple and efficient algorithm is needed,…… Why a Cell Phone:  Why a Cell Phone Everyone has one A Smart phone is a capable computer But, it has limited resources, compared to a desktop, And, has NO float point processor A simple and efficient algorithm is needed,…… Why a Cell Phone:  Why a Cell Phone Everyone has one A Smart phone is a capable computer But, it has limited resources, compared to a desktop, And, has NO float point processor A simple and efficient algorithm is needed,…… Why a Cell Phone:  Why a Cell Phone Everyone has one A Smart phone is a capable computer But, it has limited resources, compared to a desktop, And, has NO float point processor A simple and efficient algorithm is needed,…… Our Algorithm: Feature Selection and Factor Graph:  Our Algorithm: Feature Selection and Factor Graph Feature Selection Finding simple edges of four orientation Build up the more complex features Graph Model Group the features into factors Build the data driven graph model Using a simplified version of graph model to solve it non-iteratively Our Algorithm: Feature Selection and Factor Graph:  Our Algorithm: Feature Selection and Factor Graph Feature Selection Finding simple edges of four orientations Build up the more complex features Graph Model Group the features into factors Build the data driven graph model Using a simplified version of graph model to solve it non-iteratively Our Algorithm: Feature Selection and Factor Graph:  Our Algorithm: Feature Selection and Factor Graph Feature Selection Finding simple edges of four orientations Build up the more complex features Graph Model Group the features into factors Build the data driven graph model Using a simplified version of graph model to solve it non-iteratively Feature Selection:  Feature Selection Machine learning approaches Filter banks Sliding windows Learn the filters statistically Problems Lots of filters, difficult to understand and improve Require huge training data set Possible over-learning Possibly learning wrong things Feature Selection:  Feature Selection Machine learning approaches Filter banks Sliding windows Learn the filters statistically Problems Lots of filters, difficult to understand and improve Require huge training data set Possible over-learning Possibly learning wrong things Slide15:  Feature Selection: bottom up Matched Vertical/Horizontal Edgelets:  Matched Vertical/Horizontal Edgelets Anchored Vertical Edgelets:  Anchored Vertical Edgelets --- Many factors are found according to the tops and the bottoms of the edgelets A factor graph model:  A factor graph model --- Variables --- Factors The Equations:  The Equations Our Simplifications:  Our Simplifications The variables have only two states: x=0 for ground, and x=1 for figure A factor potential function g(x) has only two values: g(x) = g1 x=1 for all x’s = g0, otherwise g1 >= g0, i.e., “if you don’t have nice things to say, say something neutral” Our Simplifications:  Our Simplifications The variables have only two states: x=0 for ground, and x=1 for figure A factor potential function g(x) has only two values: g(x) = g1 x=1 for all x’s = g0, otherwise g1 >= g0, i.e., “if you don’t have nice things to say, say something neutral” Our Simplifications:  Our Simplifications The variables have only two states: x=0 for ground, and x=1 for figure A factor potential function g(x) has only two values: g(x) = g1 x=1 for all x’s = g0, otherwise g1 >= g0, i.e., “If you don’t have anything nice to say, don’t say anything” An Example Factor Graph:  An Example Factor Graph fA(x1,x2,x5) – top of chars w/ ascender fB(x3,x4,x6,x7,x8) – top of w/o ascender fC(x3,x4,x5,x6,x7,x8) – bottoms of 1st 6 sticks fD(x4,x5,x6,x7,x8,x9) – bottoms of next 6 sticks An Example Factor Graph:  An Example Factor Graph fA(x1,x2,x5) – top of chars w/ ascender fB(x3,x4,x6,x7,x8) – top of w/o ascender fC(x3,x4,x5,x6,x7,x8) – bottoms of 1st 6 sticks fD(x4,x5,x6,x7,x8,x9) – bottoms of next 6 sticks Our Non-Iterative Algorithm:  Our Non-Iterative Algorithm Set mx->f(x=0) = 0 initially Shift mf->x(x) after each iteration such that mf->x(x=0) = 0 The Algorithm converges after first iteration It can be implemented using fixed point operation only Our Non-Iterative Algorithm:  Our Non-Iterative Algorithm Set mx->f(x=0) = 0 initially Shift mf->x(x) after each iteration such that mf->x(x=0) = 0 The Algorithm converges after first iteration It can be implemented using fixed point operation only An example:  An example More examples:  More examples Summary / Discussion:  Summary / Discussion A simplified factor graph algorithm is developed for camera cell phones Variables take only two states Factor functions g(x) take only two values “No bad mouth”: g1 >= g0 Non-iterative No floating point computation Summary / Discussion:  Summary / Discussion A simplified factor graph algorithm is developed for camera cell phones Variables take only two states Factor functions g(x) take only two values “No bad mouth”: g1 >= g0 Non-iterative No floating point computation Summary / Discussion:  Summary / Discussion A simplified factor graph algorithm is developed for camera cell phones Variables take only two states Factor functions g(x) take only two values “No bad mouth”: g1 >= g0 Non-iterative No floating point computation Summary / Discussion:  Summary / Discussion A simplified factor graph algorithm is developed for camera cell phones Variables take only two states Factor functions g(x) take only two values “No bad mouth”: g1 >= g0 Non-iterative No floating point computation Summary / Discussion:  Summary / Discussion A simplified factor graph algorithm is developed for camera cell phones Variables take only two states Factor functions g(x) take only two values “No bad mouth”: g1 >= g0 Non-iterative No floating point computation Summary / Discussion:  Summary / Discussion A simplified factor graph algorithm is developed for camera cell phones Variables take only two states Factor functions g(x) take only two values “No bad mouth”: g1 >= g0 Non-iterative No floating point computation

Related presentations


Other presentations created by Urania

rocks ppt
14. 01. 2008
0 views

rocks ppt

Haydn Shaw presentation
09. 01. 2008
0 views

Haydn Shaw presentation

Cosmetica presentacion gris
10. 01. 2008
0 views

Cosmetica presentacion gris

what is femp 5B1 5D
11. 01. 2008
0 views

what is femp 5B1 5D

private sector
12. 01. 2008
0 views

private sector

S06 Lec5a rocksmet
14. 01. 2008
0 views

S06 Lec5a rocksmet

ma dissertation
14. 01. 2008
0 views

ma dissertation

Chapter 6 day 3
14. 01. 2008
0 views

Chapter 6 day 3

chapt 3 power presentation
14. 01. 2008
0 views

chapt 3 power presentation

Music 2 Powerpoint
15. 01. 2008
0 views

Music 2 Powerpoint

suips
17. 01. 2008
0 views

suips

lead awareness
18. 01. 2008
0 views

lead awareness

4 Rafi Ahmad
19. 01. 2008
0 views

4 Rafi Ahmad

weathering
20. 01. 2008
0 views

weathering

wind
21. 01. 2008
0 views

wind

TheAfterlife
15. 01. 2008
0 views

TheAfterlife

Rahman PPT
23. 01. 2008
0 views

Rahman PPT

biotech
24. 01. 2008
0 views

biotech

SBS UAE Presentation
04. 02. 2008
0 views

SBS UAE Presentation

unit 7
05. 02. 2008
0 views

unit 7

a3
08. 02. 2008
0 views

a3

Color seminar
11. 02. 2008
0 views

Color seminar

Life on a cotton plantation
25. 01. 2008
0 views

Life on a cotton plantation

17xsl
15. 01. 2008
0 views

17xsl

VMB SEMIPLENARY SHEFFIELD 2006
16. 01. 2008
0 views

VMB SEMIPLENARY SHEFFIELD 2006

OB ch03
29. 01. 2008
0 views

OB ch03

The Rosary Prayers
29. 01. 2008
0 views

The Rosary Prayers

Ogilvy Presentation
06. 02. 2008
0 views

Ogilvy Presentation

DENA ozone
20. 02. 2008
0 views

DENA ozone

MAPLD2004 Paper198
28. 02. 2008
0 views

MAPLD2004 Paper198

EdgarSpresentation
25. 01. 2008
0 views

EdgarSpresentation

Manzotti
29. 02. 2008
0 views

Manzotti

Peter Loveland Presentation
17. 01. 2008
0 views

Peter Loveland Presentation

PARQUES TEMÁTICOS
19. 03. 2008
0 views

PARQUES TEMÁTICOS

ae bg sec1
20. 03. 2008
0 views

ae bg sec1

edwards
07. 02. 2008
0 views

edwards

c1 3
24. 03. 2008
0 views

c1 3

ecology collab chapter 5 lesson
03. 04. 2008
0 views

ecology collab chapter 5 lesson

Andean Civilizations
13. 02. 2008
0 views

Andean Civilizations

GEOL3026 3
28. 03. 2008
0 views

GEOL3026 3

Realization of Olympism DR CHANG
14. 04. 2008
0 views

Realization of Olympism DR CHANG

182AWHRETraining
17. 04. 2008
0 views

182AWHRETraining

japan60s
21. 04. 2008
0 views

japan60s

employ scs place shaping 110907
22. 04. 2008
0 views

employ scs place shaping 110907

articles 71023 2
24. 04. 2008
0 views

articles 71023 2

2008 PS Summer
10. 03. 2008
0 views

2008 PS Summer

Michael Payne IOC
30. 04. 2008
0 views

Michael Payne IOC

WLOCHY ETAP 17 04 2007
02. 05. 2008
0 views

WLOCHY ETAP 17 04 2007

JTB260401Singapore2
27. 03. 2008
0 views

JTB260401Singapore2

pps 326
14. 02. 2008
0 views

pps 326

IsiorhoHawii2Group
05. 02. 2008
0 views

IsiorhoHawii2Group

29 Lectures PPT
12. 02. 2008
0 views

29 Lectures PPT

procanal
20. 02. 2008
0 views

procanal

WreckedCenters Philly
15. 03. 2008
0 views

WreckedCenters Philly

Brain sportsconcussion
23. 01. 2008
0 views

Brain sportsconcussion

Semeria history pool
18. 01. 2008
0 views

Semeria history pool

royster
09. 01. 2008
0 views

royster

IT Scenario Orissa
31. 01. 2008
0 views

IT Scenario Orissa

HST353 15
14. 02. 2008
0 views

HST353 15

marshall 25june2007
23. 01. 2008
0 views

marshall 25june2007

Rachel Ellaway Virtual Patients
04. 02. 2008
0 views

Rachel Ellaway Virtual Patients

ep050320m
08. 04. 2008
0 views

ep050320m

heathers green cleaners training
22. 01. 2008
0 views

heathers green cleaners training

Kishan
24. 01. 2008
0 views

Kishan