GPU Approach for Handwritten Devanagari Document Binarization

The optical character recognition (OCR) is the process of converting scanned images of machine printed or handwritten text, numerals, letters, and symbols into a computer processable format such as ASCII. For creating OCR’s paperless application, a system of high speed and of better accuracy is required. Parallelization of algorithm using graphics processing unit (GPU) along with CPU can be used to speed up the processing. In GPU computing, the compute-intensive operations are performed on GPU while serial code still runs on CPU. Binarization is one of the most fundamental preprocessing techniques in the area of image processing and pattern recognition. This paper proposes an adaptive threshold binarization algorithm for GPU. The aim of this research work is to speed up binarization process that eventually will help to accelerate the processing of document recognition. The algorithm implementation is done using Compute Unified Device Architecture (CUDA) software interface by NVIDIA. An average speedup of 2× is achieved on GPU GeForce 210 having 16 CUDA cores and 1.2 compute level, over the serial implementation.

[1]  Rüdiger Westermann,et al.  Linear algebra operators for GPU implementation of numerical algorithms , 2003, SIGGRAPH Courses.

[2]  Henry S. Baird,et al.  Whole-Book Recognition , 2012, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[3]  Sandhya Arora,et al.  Studies on some Soft Computing Techniques: A Case Study for Constrained Handwritten Devnagari Characters and Numerals , 2013 .

[4]  A. James Stewart,et al.  General Calculations using Graphics Hardware with Applications to Interactive Caustics , 2000, Rendering Techniques.

[5]  Brij Mohan Singh,et al.  Parallel Implementation of Devanagari Text Line and Word Segmentation Approach on GPU , 2011 .

[6]  Thierry Géraud,et al.  Efficient multiscale Sauvola’s binarization , 2013, International Journal on Document Analysis and Recognition (IJDAR).

[7]  David K. McAllister,et al.  Fast Matrix Multiplies Using Graphics Hardware , 2001, ACM/IEEE SC 2001 Conference (SC'01).

[8]  M. Valizadeh,et al.  Degraded document image binarization based on combination of two complementary algorithms , 2009, 2009 International Conference on Advances in Computational Tools for Engineering Applications.

[9]  Bidyut Baran Chaudhuri,et al.  Indian script character recognition: a survey , 2004, Pattern Recognit..

[10]  Keechul Jung,et al.  Neural network-based text location in color images , 2001, Pattern Recognit. Lett..

[11]  Fumitaka Kimura,et al.  Recognition of Off-Line Handwritten Devnagari Characters Using Quadratic Classifier , 2006, ICVGIP.

[12]  Erik G. Learned-Miller,et al.  Improving state-of-the-art OCR through high-precision document-specific modeling , 2010, 2010 IEEE Computer Society Conference on Computer Vision and Pattern Recognition.

[13]  Mahantapas Kundu,et al.  Recognition of Non-Compound Handwritten Devnagari Characters using a Combination of MLP and Minimum Edit Distance , 2010, ArXiv.

[14]  Véronique Eglin,et al.  A new mixed binarization method used in a real time application of automatic business document and postal mail sorting , 2013, Int. Arab J. Inf. Technol..

[15]  Ishwar K. Sethi,et al.  Machine recognition of constrained hand printed devanagari , 1977, Pattern Recognit..

[16]  Keechul Jung,et al.  GPU implementation of neural networks , 2004, Pattern Recognit..

[17]  J. George Cherian Panappally,et al.  Design of graphics processing unit for image processing , 2014, 2014 First International Conference on Computational Systems and Communications (ICCSC).

[18]  Rae-Hong Park,et al.  Document image binarization based on topographic analysis using a water flow model , 2002, Pattern Recognit..

[19]  Ioannis Pratikakis,et al.  Adaptive degraded document image binarization , 2006, Pattern Recognit..

[20]  Shanq-Jang Ruan,et al.  An efficient thresholding algorithm for degraded document images based on intelligent block detection , 2008, 2008 IEEE International Conference on Systems, Man and Cybernetics.

[21]  Matti Pietikäinen,et al.  Adaptive document image binarization , 2000, Pattern Recognit..

[22]  Slav Petrov,et al.  Efficient parallel CKY parsing using GPUs , 2014, J. Log. Comput..

[23]  Patrice Y. Simard,et al.  Using GPUs for machine learning algorithms , 2005, Eighth International Conference on Document Analysis and Recognition (ICDAR'05).

[24]  Santanu Chaudhury,et al.  Devnagari numeral recognition by combining decision of multiple connectionist classifiers , 2002 .

[25]  Mohamed Akil,et al.  GPU parallel implementation of the new hybrid binarization based on Kmeans method (HBK) , 2018, Journal of Real-Time Image Processing.