Computing Reviews
Today's Issue Hot Topics Search Browse Recommended My Account Log In
Review Help
Search
Symbol recognition using spatial relations
Santosh K., Lamiroy B., Wendling L. Pattern Recognition Letters33 (3):331-341,2012.Type:Article
Date Reviewed: Jan 14 2013

As a core module of graphical document image analysis, symbol recognition has generated rich literature aiming to localize and recognize graphical symbols in different applications such as recognition and interpretation of circuit diagrams, engineering drawings, and maps. In general, there are three ways to represent symbols: statistically, structurally, and a hybrid of those two. This paper uses structural representation for the selected task. To describe the spatial relation between decomposed parts of a symbol, the authors first define a set of such parts as a vocabulary: thick, circle, corner, and extremity. Next, for each pair of such parts, they compute a reference point and build a directional relation by constructing two percentage histograms of them. Finally, the directional relation is represented in the attributed relation graph (ARG) for recognition. The matching score is a combination of the alignment score of decomposed parts and the editing score of ARG nodes. Thus, the recognition results can be sorted by ranking the matching scores between a query symbol and the labeled samples.

The experimental evaluation is based on retrieval efficiency rather than the traditional precision and recall, because when the ranking parameter K grows, the retrieval efficiency does not degenerate. Using an industrial dataset of aircraft electrical wiring diagrams, the authors report performance gains for their approach compared to basic approaches based on relation models, global signals, or pixels. In addition, the authors compared the runtime performance of the proposed method to several existing alternatives and observed some speed-up. In general, I consider this a solid piece of work, although I would also be interested in seeing some experiments on noisy input.

Reviewer:  Jin Chen Review #: CR140829 (1305-0420)
Bookmark and Share
 
Pattern Recognition (I.5 )
 
Would you recommend this review?
yes
no
Other reviews under "Pattern Recognition": Date
Classification and learning using genetic algorithms: applications in bioinformatics and Web intelligence (Natural Computing Series)
Bandyopadhyay S., Pal S., Springer-Verlag New York, Inc., Secaucus, NJ, 2007.  311, Type: Book (9783540496069), Reviews: (1 of 2)
Oct 24 2007
Classification and learning using genetic algorithms: applications in bioinformatics and Web intelligence (Natural Computing Series)
Bandyopadhyay S., Pal S., Springer-Verlag New York, Inc., Secaucus, NJ, 2007.  311, Type: Book (9783540496069), Reviews: (2 of 2)
Feb 8 2008
Computational intelligence: concepts to implementations
Eberhart R., Shi Y., Morgan Kaufmann Publishers Inc., San Francisco, CA, 2007.  496, Type: Book (9781558607590)
Feb 22 2010
more...

E-Mail This Printer-Friendly
Send Your Comments
Contact Us
Reproduction in whole or in part without permission is prohibited.   Copyright 1999-2024 ThinkLoud®
Terms of Use
| Privacy Policy