An algorithm for foreground-background separation in low quality patrimonial document images

Authors:
Carlos A. B. Mello
Affiliations:
Department of Computing Systems, University of Pernambuco, Recife, Brazil
Venue:
CIARP'07 Proceedings of the Congress on pattern recognition 12th Iberoamerican conference on Progress in pattern recognition, image analysis and applications
Year:
2007

Citing 6
Cited 0

Restoration of Archival Documents Using a Wavelet Technique

IEEE Transactions on Pattern Analysis and Machine Intelligence
Separating Text and Background in Degraded Document Images " A Comparison of Global Threshholding Techniques for Multi-Stage Threshholding

IWFHR '02 Proceedings of the Eighth International Workshop on Frontiers in Handwriting Recognition (IWFHR'02)
Historical Document Image Enhancement Using Background Light Intensity Normalization

ICPR '04 Proceedings of the Pattern Recognition, 17th International Conference on (ICPR'04) Volume 1 - Volume 01
Separating Lines of Text in Free-Form Handwritten Historical Documents

DIAL '06 Proceedings of the Second International Conference on Document Image Analysis for Libraries
Improving the Quality of Degraded Document Images

DIAL '06 Proceedings of the Second International Conference on Document Image Analysis for Libraries
Flexible Text Recovery from Degraded Typewritten Historical Documents

ICPR '06 Proceedings of the 18th International Conference on Pattern Recognition - Volume 02

Quantified Score

Hi-index	0.00

Visualization

Abstract

In this article, we present a new algorithm to deal with foreground-background separation in very degraded documents. In particular, our work is applied to patrimonial document images which suffer from several types of degradation as aging effects, noise, back-to-front ink interference, etc. Our main objective is to correctly classify ink and paper to allow an efficient segmentation of the image creating high quality monochromatic images. This makes easier the broadcast of these images through the Internet. The new algorithm is based on the classical Shannon definition of entropy and a generalization defined as Tsallis Entropy and it is compared to 19 well-known classical algorithms, including DjVu algorithm. It achieved the best results by analyzing precision, recall, accuracy, specificity, PSNR and MSE.