Advertisement

Mean-Shift segmentation and PDE-based nonlinear diffusion: toward a common variational framework for foreground/background document image segmentation

  • Fadoua DriraEmail author
  • Frank LeBourgeois
Original Paper

Abstract

The presence of noise in images of degraded documents limits the direct application of segmentation approaches and can lead to the presence of a number of different artifacts in the final segmented image. A possible solution is the integration of a pre-filtering step which may improve the segmentation quality through the reduction of such noise. This study demonstrated that combining the Mean-Shift clustering algorithm and the tensor-driven diffusion process into a joint iterative framework produced promising results. For instance, this framework generates segmented images with reduced edge and background artifacts when compared to results obtained after applying each method separately. This improvement is explained by the mutual interaction of global and local information, introduced, respectively, by the Mean-Shift and the anisotropic diffusion. Another point of note is that the anisotropic diffusion process smoothed images while preserving edge continuities. The convergence of this framework was defined automatically under a stopping criterion not previously defined when the diffusion process was applied alone. To obtain a fast convergence, the common framework utilizes the speedup algorithm of the Fukunaga and Hostetler Mean-Shift formulation already proposed by Lebourgeois et al. (International Conference on Document Analysis and Recognition (ICDAR), pp 52–56, 2013). This new variant of the Mean-Shift algorithm produced similar results to the original one, but ran faster due to the application of the integral volume. The first application of this framework was document ink bleed-through removal where noise is stemmed from the interference of the verso side on the recto side, thus perturbing the legibility of the original text. Other categories of images could also be subjected to the proposed framework application.

Keywords

Image segmentation Mean-Shift clustering algorithm Integral volume Tensor-driven diffusion Joint/iterative framework Document ink bleed-through removal 

References

  1. 1.
    Barash, D., Comaniciu, D.: A common framework for nonlinear diffusion, adaptive smoothing, bilateral filtering and mean-shift. J. Image Vis. Comput. 22(1), 73–81 (2004)Google Scholar
  2. 2.
    Beevi, S.Z., Sathik, M.M.: A robust segmentation approach for noisy medical images using fuzzy clustering with spatial probability. Eur. J. Sci. Res. 41, 437–451 (2010)Google Scholar
  3. 3.
    Bregman, L.: The relaxation method of finding the common points of convex sets and its application to the solution of problems in convex optimization. USSR Comput. Math. Math. Phys. 7(3), 200–217 (1967). doi: 10.1016/0041-5553(67)90040-7 MathSciNetCrossRefGoogle Scholar
  4. 4.
    Cheng, Y. : Mean shift, mode seeking, and clustering. IEEE Trans. Pattern Anal. Mach. Intell. 17(8), 790–799 (1995)Google Scholar
  5. 5.
    Comaniciu, D.: An algorithm for data-driven bandwidth selection. IEEE Trans. Pattern Anal. Mach. Intell. (PAMI) 25(2), 281–288 (2003)CrossRefGoogle Scholar
  6. 6.
    Comaniciu, D., Meer, P.: Mean-shift: a robust approach toward feature space analysis. IEEE Trans. Pattern Anal. Mach. Intell. (PAMI) 24(5), 603–619 (2002)CrossRefGoogle Scholar
  7. 7.
    Derin, H., Elliott, H.: Modeling and segmentation of noisy and textured images using Gibbs random fields. IEEE Trans. Pattern Anal. Mach. Intell. (PAMI) 9, 39–55 (1987)CrossRefGoogle Scholar
  8. 8.
    Dong, G., Xie, M.: Color clustering and learning for image segmentation based on neural networks. IEEE Trans. Neural Netw. 16, 925–936 (2005)CrossRefGoogle Scholar
  9. 9.
    Drira, F.: Towards restoring historic documents degraded over time. In: IEEE International Conference on Document Image Analysis for Libraries (DIAL2006), pp. 350–357 (2006). doi: 10.1109/dial.2006.43
  10. 10.
    Drira, F., LeBourgeois, F., Emptoz, H.: Restoring ink bleed-through degraded document images using a recursive unsupervised classification technique. In: IAPR Workshop on Document Analysis Systems (DAS), 3872, pp. 38–49 (2006)Google Scholar
  11. 11.
    Drira, F., Lebourgeois, F., Emptoz, H.: Restoring ink bleed-through degraded document images using a recursive unsupervised classification technique. In: Document Analysis Systems (DAS), pp. 38–49 (2006)Google Scholar
  12. 12.
    Drira, F., Lebourgeois, F., Emptoz, H.: A modified mean-shift algorithm for efficient document image restoration. In: Signal Processing for Image Enhancement and Multimedia Processing, pp. 13–25 (2008)Google Scholar
  13. 13.
    Drira, F., Lebourgeois, F., Emptoz, H.: Document images restoration by a new tensor based diffusion process: application to the recognition of old printed documents. In: 10th International Conference on Document Analysis and Recognition (ICDAR), vol. 17, pp. 321–325 (2009)Google Scholar
  14. 14.
    Drira, F., Lebourgeois, F., Emptoz, H.: A new PDE-based approach for singularity-preserving regularization: application to degraded characters restoration. Int. J. Doc. Anal. Recognit. (IJDAR) 15, 183–212 (2012). doi: 10.1007/s10032-011-0165-5 CrossRefGoogle Scholar
  15. 15.
    Freedman, D., Kisilev, P., Haifa, I.: Fast mean-shift by compact density representation. In: Proceedings of Conference on Computer Vision and Pattern Recognition (CVPR), pp. 1818–1825 (2009)Google Scholar
  16. 16.
    Fukunaga, K., Hostetler, L.D.: The estimation of the gradient of a density function. IEEE Trans. Inf. Theory 21, 32–40 (1975)CrossRefzbMATHGoogle Scholar
  17. 17.
    Gatos, B., Ntirogiannis, K., Pratikakis, I.: DIBCO 2009: document image binarization contest. Int. J. Doc. Anal. Recognit. (IJDAR) 14, 35–44 (2011)CrossRefGoogle Scholar
  18. 18.
    Guo, H., Guo, P., Lu, H.: A fast mean-shift procedure with new iteration strategy and re-sampling. In: IEEE International Conference Systems, Man and Cybernetics (SMC), pp. 2385–2389 (2006)Google Scholar
  19. 19.
    Haralick, R., Shaprio, L.: Image segmentation techniques. In: Computer Vision and Graphics and Image Processing, pp. 100–132 (1985)Google Scholar
  20. 20.
    Huang, Y., Brown, M.S., Xu, D.: User-assisted ink-bleed reduction. IEEE Trans. Image Process. 19(10), 2646–2658 (2010)Google Scholar
  21. 21.
    Kimmel, R., Malladi, R., Sochen, N.: Images as embedded maps and minimal surfaces: movies, color, texture, and volumetric medical images. Int. J. Comput. Vis. 39, 111–129 (2000)CrossRefzbMATHGoogle Scholar
  22. 22.
    Lebourgeois, F., Drira, F., Gaceb, D., Duong, J.: Fast integral mean-shift: application to color segmentation of document images. In: International Conference on Document Analysis and Recognition (ICDAR) pp. 52–56 (2013)Google Scholar
  23. 23.
    Leventon, M.E., Grimson, E.L., Faugeras, O.: Statistical shape influence in geodesic active contours. In: IEEE Conference on Computer Vision and Pattern Recognition, vol. 1, pp. 1316–1323 (2000)Google Scholar
  24. 24.
    Manjunath, B.S., Simchony, T., Chellappa, R.: Stochastic and deterministic network for texture segmentation. IEEE Trans. Acoust. Speech Signal Process. 38, 1039–1049 (1990)CrossRefGoogle Scholar
  25. 25.
    Moghaddam, R.F., Cheriet, M.: EFDM: restoration of single-sided low-quality document images. In: Proceedings of International Conference on Frontiers in Handwriting Recognition (ICFHR’2008), pp. 204–209 (2008)Google Scholar
  26. 26.
    Moghaddam, R.F., Cheriet, M.: RSLDI: restoration of single-sided low-quality document images. Pattern Recognit. Spec. Issue Handwrit. Recognit. 42, 3355–3364 (2009)CrossRefzbMATHGoogle Scholar
  27. 27.
    Nikolaou, N., Papamarkos, N.: Color reduction for complex document images. Int. J. Imaging Syst. Technol. 19(1), 14–26 (2009)Google Scholar
  28. 28.
    Ntirogiannis, K., Gatos, B., Pratikakis, I.: ICFHR 2012 competition on handwritten document image binarization (H-DIBCO 2012). In: International Conference on Frontiers in Handwriting Recognition (ICFHR), pp. 817–822 (2012)Google Scholar
  29. 29.
    Ntirogiannis, K., Gatos, B., Pratikakis, I.: ICFHR2014 competition on handwritten document image binarization (H-DIBCO 2014). In: International Conference on Frontiers in Handwriting Recognition (ICFHR) pp. 809–813 (2014)Google Scholar
  30. 30.
    Pal, N.R., Pal, S.: A review on image segmentation techniques. Pattern Recognit. (PR) 26, 1277–1294 (1993)CrossRefGoogle Scholar
  31. 31.
    Paris, S., Durand, F.: A topological approach to hierarchical segmentation using mean-shift. In: Proceedings of the IEEE Conference Computer Vision and Pattern Recognition (CVPR’07), vol. 2, pp. 1–8 (2007)Google Scholar
  32. 32.
    Perona, P., Malik, J.: Scale-space and edge detection using anisotropic diffusion. IEEE Trans. Pattern Anal. Mach. Intell. (PAMI) 17, 629–639 (1990)CrossRefGoogle Scholar
  33. 33.
    Pham, D., Xu, C., Prince, J.: A survey of current methods in medical image segmentation. Annu. Rev. Biomed. Eng. 2, 315–337 (2000)CrossRefGoogle Scholar
  34. 34.
    Perona, P., Shiota, T., Malik, J.: Anisotropic diffusion. In: International Conference on Document Analysis and Recognition (ICDAR), pp. 229–254 (1994)Google Scholar
  35. 35.
    Pratikakis, I., Gatos, B., Ntirogiannis, K.: ICDAR 2011 document image binarization contest (DIBCO 2011). In: International Conference on Document Analysis and Recognition (ICDAR), pp. 1506–1510 (2011)Google Scholar
  36. 36.
    Pratikakis, I., Gatos, B., Ntirogiannis, K.: ICDAR 2013 document image binarization contest (DIBCO 2013). In: International Conference on Document Analysis and Recognition (ICDAR), pp. 1471–1476 (2013)Google Scholar
  37. 37.
    Praveena, S.M., IlaVennila, D.: Optimization fusion approach for image segmentation using k-means algorithm. IEEE Trans. Neural Netw. 7, 680–957 (2010)Google Scholar
  38. 38.
    Qua, W., Huangb, X., Jiac, Y.: Segmentation in noisy medical images using PCA model based particle filtering. In: Proceedings of the SPIE, vol. 6914 (2008)Google Scholar
  39. 39.
    Rowley-Brooke, R., Kokaram, A.: Bleed-through removal in degraded documents. In: Proceedings of SPIE 8297, Document Recognition and Retrieval, vol. XIX (2012)Google Scholar
  40. 40.
    Rowley-Brooke, R., Pitie, F., Kokaram, A.: A ground truth bleed-through document image database. In: Theory and Practice of Digital Libraries, vol. 7489, pp. 185–196 (2012)Google Scholar
  41. 41.
    Sochen, N., Kimmel, R., Malladi, R.: A geometrical framework for low level vision. IEEE Trans. Image Process. Spec. Issue PDE Based Image Process. 7, 310–318 (1998)CrossRefzbMATHGoogle Scholar
  42. 42.
    Stamatopoulos, N., Gatos, B., Perantonis, S.J.: A method for combining complementary techniques for document image segmentation. Pattern Recognit. (PR) 42, 3158–3168 (2009)CrossRefzbMATHGoogle Scholar
  43. 43.
    Tikhonov, A., Arsenin, V.: Solution of Ill-Posed Problems. Wiley, New York (1977)zbMATHGoogle Scholar
  44. 44.
    Tonazzini, A., Salerno, E., Bedini, L.: Fast correction of bleed-through distortion in grayscale documents by a blind source separation technique. Int. J. Doc. Anal. Recognit. (IJDAR) 10, 17–25 (2007)Google Scholar
  45. 45.
    Trivedi, M., Bezdek, J.: Low-level segmentation of aerial images with fuzzy clustering. IEEE Trans. Syst. Man Cybern. 164, 589–598 (1986)CrossRefGoogle Scholar
  46. 46.
    Tschumperle, D., Deriche, R.: Vector valued image regularization with pdes: a common framework for different applications. IEEE Trans. Pattern Anal. Mach. Intell. (PAMI) 27(4), 506–517 (2005)CrossRefGoogle Scholar
  47. 47.
    Wagner, R., Fisher, M.: The string to string correction problem. J. ACM (JACM) 21, 168–173 (1974)MathSciNetCrossRefzbMATHGoogle Scholar
  48. 48.
    Wang, P., Lee, D., Gray, A., Rehg, J.: Fast mean-shift with accurate and stable convergence. In: International Conference on Artificial Intelligence and Statistics (AISTATS) (2007)Google Scholar
  49. 49.
    Weickert, J.: Scale-Space Properties of Nonlinear Diffusion Filtering with a Diffusion Tensor. Report No. 110. Laboratory of Technomathematics, University of Kaiserslautern (1994)Google Scholar
  50. 50.
    Wolf, C.: Document ink bleed-through removal with two hidden markov random fields and a single observation field. IEEE Trans. Pattern Anal. Mach. Intell. (PAMI) 32, 431–447 (2010)CrossRefGoogle Scholar
  51. 51.
    Xiao, C.: Efficient mean-shift clustering using Gaussian KD-tree. Comput. Graph. Forum J. 7, 2065–2073 (2010)CrossRefGoogle Scholar
  52. 52.
    Yang, C., Duraiswami, R., DeMenthon, D., Davis, L.: Mean-shift analysis using quasi-Newton methods. In: International Conference on Image Processing (ICIP), pp. 447–450 (2003)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2017

Authors and Affiliations

  1. 1.REGIM-Lab, ENISUniversity of SfaxSfaxTunisia
  2. 2.INSA-Lyon, CNRS, UMR5205LIRIS, University of LyonLyonFrance

Personalised recommendations