Total views : 280
Binarization of Degraded Documents using Local Thresholding based on Moving Averages
Objectives: Image Binarization is the method which converts coloured and grayscale images into black and white images, black as foreground and white as background. This paper presents the new technique for binarization of degraded documents using moving averages. Methods/Statistical Analysis: Binarization of highly degraded images is very challenging task. In the proposed technique, moving averages is calculated horizontally and vertically by setting up the value of parameters. After this, local threshold value is calculated for every pixel based on moving averages within the local rectangular block both horizontally and vertically. Findings: The proposed method has been tested with large number of degraded documents and compared with existing techniques. The proposed technique makes use of the intensity of neighbouring pixels both upward and from left to right. The proposed method produce binarization results better when compared with existing Sauvola and Niblack’s Method. Applications: The proposed method produces a bianrized image which can be used in various applications like Optical character recognition, document layout analysis.
Binarization, Degraded Documents, Grayscale, Local Thresholding, Moving Averages.
- Rosenfeld A. Digital Picture Processing. 2nd (edn). Academic Press, Inc: Orlando. 1982.
- Josef K, John I. On threshold selection using clustering criteria. IEEE Transactions Systems Man Cybernetics.1985 Oct; 15(5):652-655.
- Lawrence O. Binarization and multi thresholding of document images using connectivity. CVGIP: Graphical Models and Image Processing. 1994 Nov; 56(6):494-506.
- Nobuyuki O. A threshold selection method from gray-level histograms. IEEE Transactions on Systems, Man, and Cybernetic. 1979 Jan; 9(11):62-66.
- Kapur JN, Prasanna S, Andrew KCW. A new method for gray-level picture thresholding using the entropy of the histogram. Computer Vision Graphics Image Processing.1985 Mar; 29(3):273-285.
- Josef K, John I. Minimum error thresholding. Pattern Recognition. 1986 Jan; 19(1):41-47.
- Bolan S, Shijian L, Chew LT. Binarization of historical document images using the local maximum and minimum. Proceedings of the 9th IAPR International Workshop on Document Analysis Systems, USA, 2010 Jun, 159-166.
- Mehmet S, Bu LS. Survey over image thresholding techniques and quantitative performance evaluation. Journal of Electronic Imaging. 2004 Jan; 13(1):146-165.
- Aroop M, Soumen K. Enhancement of Image Resolution by Binarization. International Journal of Computer Applications. 2010 Nov; 10(5):15-19.
- Bivind DT, Anil KJ. Goal directed evaluation of binarization methods. IEEE transactions on Pattern analysis and Machine Intelligence.1995 Dec; 17(12):1191-1201.
- Abderrahmane K, Toufik S, Halima B. Foreground-Background Separation By Feed-Forward Neural Networks In Old Manuscripts. Informatica. 2014 Feb: 38(4):329-338.
- Niblack W. An Introduction to Digital Image Processing, 2nd (edn), Strandberg Publishing Company: Denmark,1986.
- Jaakko S, Matti P. Adaptive document image binarization. Pattern Recognition. 2000 Jan; 33(1):225-236.
- Sepideh Y, Rubiyah Y, Alireza K, Mohsen P, Amirshahram H. Image Segmentation Methods and Applications in MRI Brain Images. IETE Technical Review. 2015 Jul; 32(6):413-427.
- Marian W, Ibrahima F, Dayang R. Document Image Binarization Using Retinex and Global Thresholding. Electronic Letters on Computer Vision and Image Analysis. 2015 Sep; 14(1):61-73.
- Jung GK, Nam IC, Kyoung ML. Map-Mrf Approach For Binarization of Degraded Document Image. Proceedings of the 15th IEEE International Conference on Image Processing, CA. 2008 Oct, 2612-2615.
- Khurram K, Imran S, Claudie F, Nicole V. Comparison of Niblack inspired binarization methods for ancient documents. Proceedings of the SPIE 7247, Document Recognition and Retrieval, USA. 2009 Jan, 1-9.
- Digitization. http://whatis.techtarget.com/definition/digitization. Date accessed: 04/2007.
- There are currently no refbacks.
This work is licensed under a Creative Commons Attribution 3.0 License.