Recognizing dactylogical symbols with image segmentation and a new differentiated weighting scheme

  • Authors:

  • Laura Jeanine Razo Gil;Salvador Godoy-Calderón;Ricardo Barrón Fernández

  • Affiliations:

  • Centro de Investigación en Computación, Mexico City, Mexico;Centro de Investigación en Computación, Mexico City, Mexico;Centro de Investigación en Computación, Mexico City, Mexico

  • Venue:
  • MICAI'10 Proceedings of the 9th Mexican international conference on Artificial intelligence conference on Advances in soft computing: Part II
  • Year:
  • 2010

Quantified Score

Hi-index 0.00

Visualization

Abstract

This paper contains the theoretical mechanisms of the techniques for non-convex shape recognition, through the use of contour chains and differentiated weighting scheme on them. As an application example, we use a set of digital images that represent the various symbols contained in the dactylogical alphabet. First we introduce the reader to the many pre-processing and segmentation techniques applied to the set of images. Later on, we describe the use of direction codes to code the symbols' contour. Finally, a novel differentiated weighting scheme is incorporated to an ALVOT-type algorithm, and then used for the supervised classification (identification) of the various symbols within the image set. The proposed methodology is then evaluated and contrasted through a series of experiments.