News & Politics

A Novel GA Based OCR Enhancement and Segmentation Methodology for Marathi Language in Bimodal Framework

A Novel GA Based OCR Enhancement and Segmentation Methodology for Marathi Language in Bimodal Framework
of 8
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  See discussions, stats, and author profiles for this publication at: A Novel GA Based OCR Enhancement andSegmentation Methodology for MarathiLanguage in Bimodal Framework  Chapter   in  Communications in Computer and Information Science · January 2011 DOI: 10.1007/978-3-642-19403-0_47 CITATIONS 3 READS 83 3 authors , including:Amarjot SinghUniversity of Cambridge 42   PUBLICATIONS   76   CITATIONS   SEE PROFILE Ketan BacchuwarEcole Supérieure d'Ingénieur en Electroniqu… 7   PUBLICATIONS   18   CITATIONS   SEE PROFILE All content following this page was uploaded by Ketan Bacchuwar on 21 July 2014. The user has requested enhancement of the downloaded file. All in-text references underlined in blue are added to the srcinal documentand are linked to publications on ResearchGate, letting you access and read them immediately.  C. Singh et al. (Eds.): ICISIL 2011, CCIS 139, pp. 271–277, 2011. © Springer-Verlag Berlin Heidelberg 2011 A Novel GA Based OCR Enhancement and Segmentation Methodology for Marathi Language in Bimodal Framework Amarjot Singh, Ketan Bacchuwar, and Akash Choubey Dept. of Electrical Engineering, NIT Warangal, A.P., India,, Abstract.  Automated learning systems used to extract information from images play a major role in document analysis. Optical character recognition or OCR has been widely used to automatically segment and index the documents from a wide space. Most of the methods used for OCR recognition and extraction like HMM’s, Neural etc, mentioned in literature have errors which require human operators to be rectified and fail to extract images with blur as well as illumina-tion variance. This paper explains proposes an enhancement supported thresh-old based pre-processing methodology for word spotting in Marathi printed bimodal images using image segmentation. The methodology makes use of an enhanced image obtained by histogram equalization followed by followed by age segmentation using a specific threshold. The threshold can be obtained us-ing genetic algorithms. GA based segmentation technique is codified as an optimization problem used efficiently to search maxima and minima from the histogram of the image to obtain the threshold for segmentation. The system described is capable of extracting normal as well as blurred images and images for different lighting conditions. The same inputs are tested for a standard GA based methodology and the results are compared with the proposed method. The paper further elaborates the limitations of the method. Keywords: OCR, Genetic Algorithm, Bimodal, Blur, Illumination. 1 Introduction Optical character recognition (OCR) is a mature field, born in 1950’s initially prac-ticed with electronic and electro-mechanical methods aimed towards machine printing while limited to fixed-format applications. A typical OCR system is widely used for character segmentation, segment reconstruction, character recognition, and word and phrase construction without human intervention or human correction. The basic step before all the application mentioned above is image enhancement and segmentations, hence the paper focuses on the same. According to literature survey, a number of methods have been proposed for image enhancement and segmentation of OCR but the in ability of these systems to work together for different experimental conditions is a major drawback in document  272 A. Singh, K. Bacchuwar, and A. Choubey analysis. Besides the poor quality of images, illumination changing conditions also makes OCR recognition a daunting task. The error incurred is corrected using a hu-man support system which is a time consuming as well as a costly proposal. This paper proposes an OCR system which makes use of histogram equalization to extract images blur as well as illumination varying conditions simultaneously. The histogram used by the mentioned algorithm is bimodal in nature hence it can be divided into two classes. Genetic algorithm is further used to select the threshold from the histogram for extracting the object from the background. The capabilities of the system are tested on images with blur, noise and change in illumination. The paper is divided into five sections. The next section elaborates the related work to OCR enhancement and recognition developed over the past few years. Third section explains the standard and proposed methodology used in the paper to enhance and extract the characters followed by the results obtained from the simulations in the fourth section. A brief summary of the paper is presented in the last section of the paper.   Fig. 1.  (a) The input RGB image further converted into gray scale, the results along with the histogram obtained from standard algorithm (b The input blur image, the corresponding results along with the respective histograms obtained from standard algorithm (c) The input illumina-tion variant image, the results along with the respective histograms obtained from standard algorithm 2 Related Work In case of bimodal images, the histogram has a deep and sharp valley between two peaks representing objects and back ground respectively which can be used to select the threshold representing the bottom of this valley.   Khankasikam  et. al [1], [6] pro-posed the valley sharpening techniques which restricts the histogram to the pixels with large absolute values of derivatives where as S. Watanable et. al. [2], [7] proposed the difference histogram method, which selects threshold at the gray level with the maximal amount of difference. These techniques utilize the information concerning neighboring pixels or edges in the srcinal picture to modify the histogram   A Novel GA Based OCR Enhancement and Segmentation Methodology 273 so as to make it useful for thresholding. Another class of methods deal directly with the grey level histogram by parametric techniques. The histogram is approximated in the least square sense by a sum of Gaussian distributions, and statistical decision procedures are applied [8]. However, such methods are tedious and involve high computational power. In Di Gesu [3], [8] the idea of using both intensities and spatial information has been considered to take into account local information used in human perception. A number of new strategies and methodologies have been proposed over the last couple of years to detect the global as well as local solutions in a nonlinear multimodal function optimization [4], [5], [9]. Multiple peaks can be maintained in multimodal optimization problem with the help of crowding. Crowding method is extremely efficient in detecting in detecting the two peaks on a bimodal histogram. Further, GA can be applied to discover the valley bottom between these peaks which can be used as the threshold for extracting the information from the background. 3 Algorithm The section discusses the standard GA based algorithm being used for information extraction. In the second section, a novel histogram based methodology which can be efficiently used to extract information from bimodal images is presented. The algorithms are explained in detail below. 3.1 Standard GA Based Algorithm The algorithm is a histogram based approach effectively used to extract useful infor-mation from bimodal images. The histogram of the digital image is a plot or graph of the frequency of occurrence of each gray level in the image across gray scale values. Genetic algorithm is applied on the histogram of the bimodal image to extract the useful information from the background. A random population of size  N   is initial-ized where the element acquire the value between 0 to 255. The crossover and muta-tion operations are carried out on the randomly chosen two parents. Appropriate value of crossover probability ( c P ) and mutation probability ( m P ) is fixed. The winner of each tournament (the one with the best fitness) is selected. After computing the fitness value of the off-springs, Tournament selection strategy is used to allow off- springs to compete with the parents. It involves running a competition among two individuals Fig. 2.  (a) The input image and the corresponding enhanced image (b) The equalized histogram of the resultant enhanced image    274 A. Singh, K. Bacchuwar, and A. Choubey chosen at random from the population. The fittest between both is selected. This is the method used in Genetic algorithm for selection of individual from the population. Here two parents and two-off springs compete to give two best individuals as result. The resulting selected elements are located in their respective classes. The methodol-ogy used is termed as Crowding Method. This method basically replaces the older elements in the population by the fittest elements in the resulting generation which helps to reduce replacement error. The repetition performed for all the elements re-sults into convergence. This converged value is the gray value corresponding to the minima between two peaks. Then this gray value is used as threshold value and the image is segmented. Fig. 3.  (a) The input, enhanced and resultant image for illumination changes for proposed algorithm (b) The input, enhanced and resultant image for blur for proposed algorithm  
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!