Texture defects detection on microscale images of materials
A.I. Plastinin, A.G. Khramov, V.A Soifer

Full text of article: Russian language.

Abstract:
The article presents a method for texture defect detection based on image neighborhoods set analysis. Method allows estimating decision function based on images with defects, as well as on images without defects. We provide steel microstructure defects detection results that show the advantages of described method.

Key words:
texture images, markov random fields, single-class SVM, kernel function, non-linear separation boundary.

References:

  1. Tsai, I. Applying an arti?cial neural network to pattern recognition in fabric defects / I. Tsai, C. Lin, J. Lin // Textile Research Journal. - 1995. - Vol. 65. - P. 123-130.
  2. Siew, L. Texture measures for carpet wear assessment / L. Siew, R. Hodgson, E. Wood // IEEE Transactions on Pattern Analysis and Machine Intelligence. - 1988. - Vol. 10. - P. 92-105.
  3. Iivarinen, J. Unsupervised segmentation of surface defects / J. Iivarinen, J. Rauhamaa, A. Visa // International Conference on Pattern Recognition. - 1996. - Vol. 4. - P. 356-360.
  4. Chen, J. A structural approach to identify defects in textured images / J. Chen, A. Jain // IEEE International Conference on Systems, Man, and Cybernetics. - 1988. - Vol. 1. - P. 29-32.
  5. Mallick-Goswami, B. Detecting defects in fabric with laser-based morphological image processing / B. Mallick-Goswami, A. Datta // Textile Research Journal. - 2000. - Vol. 70. - P. 758-762.
  6. Tsai, D. Automated surface inspection for directional textures / D. Tsai, C. Heish // Image and Vision Computing. - 1999. - Vol. 18, N 1. - P. 49-62.
  7. Kumar, A. Defect detection in textured materials using Gabor ?lters / A. Kumar, G. Pang // IEEE Transactions on Industry Applications. - 2002. - Vol. 38, N 2. - P. 425-440.
  8. Kim, S. Wavelet analysis to fabric defect detection in weaving processes / S. Kim, M. Lee, K. Woo // Proceedings of IEEE International Symposium on Industrial Electronics. - 1999. - Vol. 3. - P. 1406-1409.
  9. Mao, J. Texture classi?cation and segmentation using mul­tiresolution simultaneous autoregressive models / J. Mao, A. Jain // Pattern Recognition. - 1992. - Vol. 25, N 2. - P. 173-188.
  10. Conci, A. A fractal image analysis system for fabric inspection based on box-counting method / A. Conci, C. Proenca // Computational Networks. - 1998. - Vol. 30. - P. 1887-1895.
  11. Cohen, F. Automated inspection of textile fabrics using textural models / F. Cohen, Z. Fan, S. Attali // IEEE Transactions on Pattern Analysis and Machine Intelligence. - 1991. - Vol. 13, N 8. - P. 803-809.
  12. Markou, M. Novelty Detection: A Review - Part 1: Statistical Approaches / M. Markou, S. Singh // Signal Processing. - 2003. - Vol. 83. - P. 2481-2497.
  13. Markou, M. Novelty detection: a review - part 2: neural network based approaches / M. Markou, S. Singh // Signal Processing. - 2003. - Vol. 83. - P. 2499-2521.
  14. Li, Stan Z. Markov Random Field in Image Analysis / Stan Z. Li. - Springer-Verlag, 2009. - 362 p.
  15. Winkler, G. Image Analysis, Random Fields and Dyna­mic Monte Carlo Methods / G. Winkler. - Springer-Verlag, 1995. - 324 p.
  16. Efros, A.A. Texture synthesis by non-parametric sampling / A.A. Efros, T.K. Leung // ICCV. - 1999. - Vol. 2. - P. 1033-1038.
  17. Paget, R. Nonparametric Markov Random Field Models for Natural Textures Images. Ph.D. thesis. - University of Queensland, St Lucia, QLD Australia, December 1999.
  18. Pavlidis, T. Algorithms for Graphics and Image Processing / T. Pavlidis. - Computer Science Press, Maryland, 1982. - 416 p.
  19. Scholkopf, B. Learning with Kernels Support Vector Machines, Regularization, Optimization and Beyond / B. Scholkopf, A.J. Smola. - MIT Press, 2001. - 644 p.
  20. Schölkopf, B. Estimating the Support of a High-Dimen­sional Distribution / B. Schölkopf, J.C. Platt, J.S. Taylor, A.J. Smola, R.C. Williamson // Neural Computation. - 2001. - Vol. 13, N 7. - P. 1443-1471.
© 2009, IPSI RAS
Institution of Russian Academy of Sciences, Image Processing Systems Institute of RAS, Russia, 443001, Samara, Molodogvardeyskaya Street 151; E-mail: ko@smr.ru; Phones: +7 (846) 332-56-22, Fax: +7 (846) 332-56-20