Advanced search
1 file | 463.98 KB

Geometric local binary patterns a new approach to analyse texture in images

Author
Organization
Abstract
In image processing, applications such as quality inspection of surface degradation requires algorithms capable of differentiating similar textures. Exploring symmetries of local texture helps to increase the discriminance of texture algorithms. Particularly, Local Binary Pattern (LBP) techniques are highlighted in this field because their robustness against intensity variations. However, the sampling exploited by LBP operators based on symmetries cover basic neighbourhoods given by points on a circle that do not describe well in some cases the changes when the textures are similar. We present in this paper a new technique based on symmetries, called the Geometric Local Binary Pattern (GLBP) technique, that describe intensity changes using the information given by points on several circles. We compared one particular case of the GLBP technique with two configurations of the LBP technique, in which points on two circles with different radii were independently computed. We evaluated the performance of the algorithms in discriminating textures from a set composed with samples from the Brodatz texture database.
Keywords
Carpet Wear, Local Binary Patterns, Texture Inspection

Downloads

  • (...).pdf
    • full text
    • |
    • UGent only
    • |
    • PDF
    • |
    • 463.98 KB

Citation

Please use this url to cite or link to this publication:

Chicago
Orjuela Vargas, Sergio Alejandro, Rolando Augusto Quiñones Lara, Benhur Ortiz Jaramillo, Filip Rooms, Ewout Vansteenkiste, Robain De Keyser, and Wilfried Philips. 2012. “Geometric Local Binary Patterns a New Approach to Analyse Texture in Images.” In Selected Papers of the 2010 International Conference on Topology and Its Applications, ed. Dimitris N Georgiou, Stavros D Iliadis , and Ioannis E Kougias, 132–144. Patras, Greece: Technological Educational Institute (TEI) of Messolonghi.
APA
Orjuela Vargas, S. A., Quiñones Lara, R. A., Ortiz Jaramillo, B., Rooms, F., Vansteenkiste, E., De Keyser, R., & Philips, W. (2012). Geometric local binary patterns a new approach to analyse texture in images. In D. N. Georgiou, S. D. Iliadis , & I. E. Kougias (Eds.), Selected papers of the 2010 international Conference on Topology and its applications (pp. 132–144). Presented at the 2010 International Conference on Topology and its applications, Patras, Greece: Technological Educational Institute (TEI) of Messolonghi.
Vancouver
1.
Orjuela Vargas SA, Quiñones Lara RA, Ortiz Jaramillo B, Rooms F, Vansteenkiste E, De Keyser R, et al. Geometric local binary patterns a new approach to analyse texture in images. In: Georgiou DN, Iliadis SD, Kougias IE, editors. Selected papers of the 2010 international Conference on Topology and its applications. Patras, Greece: Technological Educational Institute (TEI) of Messolonghi; 2012. p. 132–44.
MLA
Orjuela Vargas, Sergio Alejandro, Rolando Augusto Quiñones Lara, Benhur Ortiz Jaramillo, et al. “Geometric Local Binary Patterns a New Approach to Analyse Texture in Images.” Selected Papers of the 2010 International Conference on Topology and Its Applications. Ed. Dimitris N Georgiou, Stavros D Iliadis , & Ioannis E Kougias. Patras, Greece: Technological Educational Institute (TEI) of Messolonghi, 2012. 132–144. Print.
@inproceedings{2072095,
  abstract     = {In image processing, applications such as quality inspection of surface degradation requires algorithms capable of differentiating similar textures. Exploring symmetries of local texture helps to increase the discriminance of texture algorithms. Particularly, Local Binary Pattern (LBP) techniques are highlighted in this field because their robustness against intensity variations. However, the sampling exploited by LBP operators based on symmetries cover basic neighbourhoods given by points on a circle that do not describe well in some cases the changes when the textures are similar. We present in this paper a new technique based on symmetries, called the Geometric Local Binary Pattern (GLBP) technique, that describe intensity changes using the information given by points on several circles. We compared one particular case of the GLBP technique with two configurations of the LBP technique, in which points on two circles with different radii were independently computed. We evaluated the performance of the algorithms in discriminating textures from a set composed with samples from the Brodatz texture database.},
  author       = {Orjuela Vargas, Sergio Alejandro and Qui{\~n}ones Lara, Rolando Augusto and Ortiz Jaramillo, Benhur and Rooms, Filip and Vansteenkiste, Ewout and De Keyser, Robain and Philips, Wilfried},
  booktitle    = {Selected papers of the 2010 international Conference on Topology and its applications},
  editor       = {Georgiou, Dimitris N and  Iliadis , Stavros D and Kougias, Ioannis E},
  keyword      = {Carpet Wear,Local Binary Patterns,Texture Inspection},
  language     = {eng},
  location     = {Nafpaktos, Greece},
  pages        = {132--144},
  publisher    = {Technological Educational Institute (TEI) of Messolonghi},
  title        = {Geometric local binary patterns a new approach to analyse texture in images},
  year         = {2012},
}