A vectorized thinning algorithm for handwritten symbols recognition

• Autorzy: Stąpor K.
• Języki publikacji: EN

Abstrakty

EN

In this paper a non-pixelwise thinning algorithm of binary line images, called vectorized thinning is proposed. The praesented algorithm produces a skeleton of a 2D-object which is performed in three steps: 1) links finding and simple region extraction, 2) complex region extraction and multiple points finding, 3) transformation into vectorized skeleton. As opposed to other thinning algorithms, the obtained skeleton is yet in vector from, particularly suitable for feature extraction in the cartographic symbol recognition from scanned, geodesic maps, with much better results than using other, pixelwise thinning methods ([17]). The main advantages of the proposed thinning algorithm lie in better extraction of multiple points representing corners, branch or crossing regions of 2D objects, and less sensitivity to a boundary noise which is one of the main problems in pixelwise thinning algorithms.

Słowa kluczowe

EN

thinning algorithm; skeleton; medial axis; feature extraction; shape analysis

• Wydawca: Institute of Information Technology of the Warsaw University of Life Sciences – SGGW
• Czasopismo: Machine Graphics and Vision
• Rocznik: 1999
• Tom: Vol. 8, No. 3
• Strony: 341--352
• Opis fizyczny: Bibliogr. 17 poz.

Twórcy

autor

Stąpor K.

• Institute of Computer Science, Silesian Technical University,