Summary
Previous formulations of the global Mean Shift clustering algorithm incorporate a global mode finding which requires a lot of computations making it extremely time-consuming. This paper focuses on reducing the computational cost in order to process large document images. We introduce thus a local-global Mean Shift based color image segmentation approach. It is a two-steps procedure carried out by updating and propagating cluster parameters using the mode seeking property of the global Mean Shift procedure. The first step consists in shifting each pixel in the image according to its R-Nearest Neighbor Colors(R-NCC) in the spatial domain. The second step process shifts only the previously extracted local modes according to the entire pixels of the image.
Our proposition has mainly three properties compared to the global Mean Shift clustering algorithm: 1) an adaptive strategy with the introduction of local constraints in each shifting process, 2) a combined feature space of both the color and the spatial information, 3) a lower computational cost by reducing the complexity. Assuming all these properties, our approach can be used for fast pre-processing of real old document images. Experimental results show its desired ability for image restoration; mainly for ink bleed-through removal, specific document image degradation.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
K. Fukunaga, Introduction to Statistical Pattern Recognition. Boston, MA: Academic Press, 1990.
F. Drira, Towards restoring historic documents degraded over time. Dans Second IEEE International Conference on Document Image Analysis for Libraries (DIAL2006), Lyon, France. pp. 350-357. ISBN 0-7695-2531-4. 2006.
Y. Cheng, Mean shift, mode seeking, and clustering. Pattern Analysis and Machine Intelligence, IEEE Transactions, vol. 17, pp. 790-799, 1995.
D. Comaniciu and P. Meer, Mean shift: A robust approach toward feature space analysis. IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 24, no. 5, pp. 603-619, 2002.
Y. Leydier, F. LeBourgeois, H. Emptoz, Serialized k-means for adaptative color image segmentation . application to document images and others. DAS 2004, LNCS 3163, Italy, September 2004, 252-263.
F. Drira, F. Lebourgeois, H. Emptoz Restoring Ink Bleed- Through Degraded Document Images Using a Recursive Unsupervised Classification Technique. DAS2006, LNCS 3872. Nelson, New Zealand, 2006, 38-49.
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2008 Springer Science+Business Media, LLC
About this chapter
Cite this chapter
Drira, F., Lebourgois, F., Emptoz, H. (2008). A Modified Mean Shift Algorithm For Efficient Document Image Restoration. In: Damiani, E., Yétongnon, K., Schelkens, P., Dipanda, A., Legrand, L., Chbeir, R. (eds) Signal Processing for Image Enhancement and Multimedia Processing. Multimedia Systems and Applications Series, vol 31. Springer, Boston, MA. https://doi.org/10.1007/978-0-387-72500-0_2
Download citation
DOI: https://doi.org/10.1007/978-0-387-72500-0_2
Publisher Name: Springer, Boston, MA
Print ISBN: 978-0-387-72499-7
Online ISBN: 978-0-387-72500-0
eBook Packages: Computer ScienceComputer Science (R0)