A graph-edit algorithm for hand-drawn graphical document recognition and thier automatic introduction into CAD systems

• Autorzy: Lladós J. , Martí E.
• Języki publikacji: EN

Abstrakty

EN

In this work, a graph-based algorithm for symbol recognition in hand-drawn architectural plans has been described. The algorithm belongs to a prototype of man-machine interface consisting in the introduction of hand-drawndesigns to a CADsystem. Documentsand symbol prototypes are represented in terms of a Region Adjecency Graph (RAG) structure. Hence, the localization of symbol instaces in documents is performed by an error-tolerant subgraph isomorphism algorithm that looks for the minimum cost edit sequence that transforms a model graph to an input one. In this paper we describe this algorithm and the set of graph edit operations designed to transform RAGs. The main idea of the algorithm is to formulate the distance between two RAGs in terms of the string edit distance between the boundary strings of the corresponding regions. The main advantage of the algorithm is its ability to cope with distorted structuredand its invariance to rotation, translation and scaling.

Słowa kluczowe

EN

graphic recognition; graph matching; edit distance; CAD systems; structural pattern recognition

• Wydawca: Institute of Information Technology of the Warsaw University of Life Sciences – SGGW
• Czasopismo: Machine Graphics and Vision
• Rocznik: 1999
• Tom: Vol. 8, No. 2
• Strony: 195--211
• Opis fizyczny: Bibliogr. 22 poz.

Twórcy

autor

Lladós J.

autor

Martí E.

• Computer Vision Center - Dept. Informatica. Univeritat Autonoma de Barcelona, 08193 Ballaterra (Barcelona), Spain