Character string recognition on maps, a rotation-invariant recognition method

Deseilligny, Marc Pierrot; Men, Hervé Le; Stamon, Georges

doi:10.1016/0167-8655(95)00084-5

Cited by 30 publications

(17 citation statements)

References 1 publication

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Recognized elements are taken from cadastral maps such as presented in Figure 4. For recognized elements translation, scale and rotation are normalized [11]. Next, low cost features such as histograms and projection are calculated.…”

Section: Testsmentioning

confidence: 99%

Distances Tree as SVM Ensemble in Digits Recognition Task

Luckner¹

2008

Proceedings of the 11th Joint Conference on Information Sciences (JCIS)

View full text Add to dashboard Cite

show abstract

Section: Testsmentioning

confidence: 99%

Distances Tree as SVM Ensemble in Digits Recognition Task

Luckner¹

2008

Proceedings of the 11th Joint Conference on Information Sciences (JCIS)

View full text Add to dashboard Cite

show abstract

“…However, some aspects of this pattern recognition problem are still the object of intensive research activities because of their specificity: multi-oriented and multi-scaled characters [20,23,42,43], connected characters [20,42,44], connected symbols [6], and geometric features identification [45]. This kind of problem is crucial, since characters and symbols carry very important information about the contents of a document: poor detection/recognition of such patterns leads to some very important problems in the interpretation process.…”

Section: Technical Documents Analysismentioning

confidence: 99%

Technical Map Interpretation: A Distributed Approach

Ogier

Mullot

Labiche

et al. 2000

Pattern Analysis & Applications

View full text Add to dashboard Cite

This paper deals with the general problem of document understanding. We propose the description of a formal architecture of a device capable of interpreting technical and cartographic documents. This device relies on two main points, i.e. a model of the document and the implementation of a set of 'builders', the aim of which is to progressively construct information of as high a semantic level as that provided by the document drawer. Two main stages are integrated in the reasoning process: the first one consists in constructing the information, through a bottom-up approach. Then, a cycling stage is triggered to solve ambiguities detected by the system and corresponding to inconsistent objects with regard to the document model. In this paper, the whole approach is presented in the context of the French cadaster interpretation. The first implementation has enabled us to quantify the interpretation results and to verify the relevance of the cycling stage.

show abstract

“…(i) their completeness: given the quantities M/(0,0),M/(0,1) and {Ij(v,q)}, it is possible to return to the whole set of FM coefficients {Mf(v,q)}, and thus to reconstruct f(p, 9), by using the inverse AFMT in the following way:…”

Section: Analytic Prolongation Of the Fourier-mellin (Afmt) And Propementioning

confidence: 99%

A Multiscale and Multiorientation Recognition Technique Applied to Document Interpretation: Application to the French Telephone Network Maps

Cariou¹,

Ogier

Adam

et al. 1999

Int. J. Patt. Recogn. Artif. Intell.

View full text Add to dashboard Cite

In this paper, we propose an original methodology which allows the detection and the recognition of multioriented and multiscaled patterns. The supports on which the method is applied are technical documents representing the network of the French Telephone operator France Telecom, overlaid on urban maps. The adopted technique, based on the Fourier-Mellin transform (FMT) is integrated in a global strategy that solves ambiguous situations, through the providing of contextual information. The strategy, which is applied to solve the character and symbol classification problem, can be divided into two stages. The first one consists in constructing a set of invariant descriptors for each isolated pattern which is extracted from a characters layer given by the system approach. The second stage, based on a filtering scheme, consists in detecting and recognizing connected shapes. The results of the application of this technique are very encouraging since the classification rate reaches excellent scores in comparison with classical techniques.

show abstract

Character string recognition on maps, a rotation-invariant recognition method

Cited by 30 publications

References 1 publication

Distances Tree as SVM Ensemble in Digits Recognition Task

Distances Tree as SVM Ensemble in Digits Recognition Task

Technical Map Interpretation: A Distributed Approach

A Multiscale and Multiorientation Recognition Technique Applied to Document Interpretation: Application to the French Telephone Network Maps

Contact Info

Product

Resources

About