A new Tsallis entropy-based thresholding algorithm for images of historical documents

ABSTRACT

It is presented in this paper an algorithm for thresholding images of historical documents. The main objective is to generate high quality monochromatic images in order to make them easily accessible thru Internet and achieve high recognition rates by Optical Character Recognition algorithms. Our new algorithm is based on the classical entropy concept and a variation defined by the Tsallis Entropy and it proved to be more efficient than classical thresholding algorithms. The images generated are analyzed using precision, recall, accuracy and specificity.

REFERENCES

• 1: C. A. B. Mello et al.: Image Thresholding of Historical Documents: Application to the Joaquim Nabuco's File, Digital Cultural Heritage Conference - Eva Vienna (2006), p. 115--122, Austria.
• 2: C. A. B. Mello, A. H. M. Costa: Image Thresholding of Historical Documents Using Entropy and ROC Curves, Lec. Notes in Computer Science (2005), v. 3773, p. 905--916.
• 3: A. L. I. Oliveira , C. A. B. Mello , E. R. Silva Jr , V. M. O. Alves Alves, Optical Digit Recognition for Images of Handwritten Historical Documents, Proceedings of the Ninth Brazilian Symposium on Neural Networks, p.29, October 23-27, 2006 [doi>10.1109/SBRN.2006.34]
• 4: J. R. Parker: Algorithms for Image Processing and Computer Vision. John Wiley & Sons (1997).
• 5: C. Shannon: A Mathematical Theory of Communication. Bell System Technology Journal (1948), vol. 27, pp. 370--423, 623--656.
• 6: C. Tsallis: Possible Generalization of Boltzmann-Gibbs statistics, J. of Statistical Physics (1988), vol. 52, nos. 1-2, pp. 479--487.
• 7: J. N. Kapur: Measures of Information and their Applications, J. Wiley & Sons (1994).
• 8: M. Sezgin, B. Sankur. "Survey over image thresholding techniques and quantitative performance evaluation", J. of Electronic Imaging, no.13, vol 1, pp. 146--165, 2004.
• 9: João Marcelo M. da Silva , Rafael Dueire Lins , Valdemar Cardoso da Rocha, Binarizing and filtering historical documents with back-to-front interference, Proceedings of the 2006 ACM symposium on Applied computing, April 23-27, 2006, Dijon, France [doi>10.1145/1141277.1141471]

INDEX TERMS

Primary Classification:
  I. Computing Methodologies
  I.4 IMAGE PROCESSING AND COMPUTER VISION
      I.4.6 Segmentation
          Subjects: Pixel classification

General Terms:
Algorithms, Experimentation

Keywords:
document processing, entropy, historical documents, image segmentation, thresholding