Symbol Recognition by Global Local Structural Approaches, Based on Scenario Use, and with a XML Representation of Data - PowerPoint PPT Presentation

About This Presentation
Title:

Symbol Recognition by Global Local Structural Approaches, Based on Scenario Use, and with a XML Representation of Data

Description:

circle erasing images subtraction strategy 1. ICISP'2003 : Agadir, Morocco : 24 June, 2003 ... circle erasing skeleton graph subgraph. room detection ... – PowerPoint PPT presentation

Number of Views:36
Avg rating:3.0/5.0
Slides: 34
Provided by: mathieuDe8
Category:

less

Transcript and Presenter's Notes

Title: Symbol Recognition by Global Local Structural Approaches, Based on Scenario Use, and with a XML Representation of Data


1
Symbol Recognition by Global Local Structural
Approaches,Based on Scenario Use,and with a XML
Representation of Data
  • Mathieu Delalandre¹, Stéphane Nicolas¹, Eric
    Trupin¹, Jean-Marc Ogier²
  • ¹PSI Laboratory, Rouen University, France
  • ²L3I Laboratory, La Rochelle University, France
  • Phd thesis website http//site.voila.fr/roxml

2
Plan
  • Structural Shape Recognition
  • Systems Part I, Graph-Matching
  • Systems Part II, Global Analysis
  • Systems Part III, Local Analysis
  • Scenarios Construction and Use
  • Experiments and Results
  • XML Use in System
  • Conclusion and Perspectives

3
Plan
  • Structural Shape Recognition
  • Systems Part I, Graph-Matching
  • Systems Part II, Global Analysis
  • Systems Part III, Local Analysis
  • Scenarios Construction and Use
  • Experiments and Results
  • XML Use in System
  • Conclusion and Perspectives

4
Structural Shape Recognition
  • structural shape recognition, two steps
  • structural analysis (Delalandre 2003)
  • structural recognition (Tombre 1996)

Structural Analysis
Structural Recognition
ltsymbolsgt ltlabelsgtlt/labelsgt lt/symbolsgt
5
Structural Shape Recognition
  • applications
  • oriental handwriting (Lau 2002)
  • structured document (Naguy 2000)
  • graphical document (Song 2003)
  • symbol structured and graphical document
    (Lladós 2001)

6
Structural Shape Recognition
  • structural recognition
  • graph-matching approach (Hancock 2002)
  • grammar approach (Blostein 1995)
  • structural analysis (Delalandre 2003)
  • global local approaches
  • ? multi representation
  • ? process simplification

7
Plan
  • Structural Shape Recognition
  • Systems Part I, Graph-Matching
  • Systems Part II, Global Analysis
  • Systems Part III, Local Analysis
  • Scenarios Construction and Use
  • Experiments and Results
  • XML Use in System
  • Conclusion and Perspectives

8
Systems Part I,Graph-Matching
  • exact and inexact graph algorithms
  • (Héroux 2000)
  • exact and inexact subgraph algorithms
  • Ullmans method (Ullman 1976)
  • clique based detection method
  • exact VF methods (Foggia 2001)
  • exact and inexact Messmers decomposition methods
    (Messmer 1995)

9
Plan
  • Structural Shape Recognition
  • Systems Part I, Graph-Matching
  • Systems Part II, Global Analysis
  • Systems Part III, Local Analysis
  • Scenarios Construction and Use
  • Experiments and Results
  • XML Use in System
  • Conclusion and Perspectives

10
Systems Part II,Global Analysis
  • global analysis based on statistical recognition
    result use (Delalandre 2002)

11
Systems Part II,Global Analysis
loop extraction
(1)
connected component filtering
(2)
feature extraction
(3)
statistical classification
(4)
graphs
image processings
global analysis
(5)
recognition processings
12
Systems Part II,Global Analysis
  • (1) loop extraction, (2) connected component
    filtering

13
Systems Part II,Global Analysis
  • (3) feature extraction, (4) statistical
    recognition

14
Systems Part II,Global Analysis
  • (5) global analysis
  • distance and/or closeness constraints
  • ? grouping or segmentation strategy

15
Plan
  • Structural Shape Recognition
  • Systems Part I, Graph-Matching
  • Systems Part II, Global Analysis
  • Systems Part III, Local Analysis
  • Scenarios Construction and Use
  • Experiments and Results
  • XML Use in System
  • Conclusion and Perspectives

16
Systems Part III,Local Analysis
  • vectorisation scenario, three levels

17
Systems Part III,Local Analysis
  • image level
  • a) skeletonisation (Di Baja 1992) contouring
  • b) skeletonisation adaptation

18
Systems Part III,Local Analysis
  • boundary level
  • a) pixel list extraction based on 3-connected
    destruction
  • b) object simplification (Song 2002)

19
Systems Part III,Local Analysis
  • structured data level (1)
  • a) interiority segmentation
  • b) polygonisation (Ramer 1972) (Pavlidis 1986),
    and circle fitting
  • c) curve fitting (Liao 1990)

20
Systems Part III,Local Analysis
  • structured data level (2)
  • (a-b) node detection
  • (b-c) Pruning and (c-d) Merging

21
Plan
  • Structural Shape Recognition
  • Systems Part I, Graph-Matching
  • Systems Part II, Global Analysis
  • Systems Part III, Local Analysis
  • Scenarios Construction and Use
  • Experiments and Results
  • XML Use in System
  • Conclusion and Perspectives

22
Scenario Construction and Use
  • system based on scenario use (Saïdali 2003)
  • interface package
  • scenario construction and execution
  • statistical learning and structural model
    acquisition
  • Knowledge base
  • training sets
  • scenarios
  • entry data
  • processings library
  • image processings local and global
  • classification processings

23
Plan
  • Structural Shape Recognition
  • Systems Part I, Graph-Matching
  • Systems Part II, Global Analysis
  • Systems Part III, Local Analysis
  • Scenarios Construction and Use
  • Experiments and Results
  • XML Use in System
  • Conclusion and Perspectives

24
Experiments and Results
  • symbol recognition on technical drawing
  • 65 scenarios different strategies and parameter
    settings
  • 700 symbol graphs

25
Experiments and Results
  • global strategies (1, 2, 3)

strategy 1 gives good result, but dont deal the
nearest symbols
26
Experiments and Results
  • global local strategy (4)
  • circle erasing images subtraction strategy 1

27
Experiments and Results
Strategy 4 gives good result and deals the
nearest symbols
28
Experiments and Results
  • local strategy (5)
  • circle erasing skeleton graph subgraph
  • ? room detection
  • ? connected characters detection

29
Plan
  • Structural Shape Recognition
  • Systems Part I, Graph-Matching
  • Systems Part II, Global Analysis
  • Systems Part III, Local Analysis
  • Scenarios Construction and Use
  • Experiments and Results
  • XML Use in System
  • Conclusion and Perspectives

30
XML Use in System
  • sub-language use
  • SVG for graphics representations
  • XGMML for graph library
  • transformation and request use
  • XSLT to transform
  • Quilt to request

31
Plan
  • Structural Shape Recognition
  • Systems Part I, Graph-Matching
  • Systems Part II, Global Analysis
  • Systems Part III, Local Analysis
  • Scenarios Construction and Use
  • Experiments and Results
  • XML Use in System
  • Conclusion and Perspectives

32
Conclusion and Perspectives
  • Conclusion
  • local approach helps global approach
  • scenario use permits systems adaptation and
    strategy test
  • XML use permits an easy data manipulation
  • Perspectives
  • local approach (Delalandre 2003)
  • result coherency test with RuleML
  • Knowledge base structuring with RDF

33
End
  • sorry for my absent, any question ?
  • Mail mathieu.delalandre_at_univ-rouen.fr
  • Phd Website http//site.voila.fr/roxml
Write a Comment
User Comments (0)
About PowerShow.com