Informatics Publishing Limited

R. Indra Gandhi ¹, K. Iyakutti ²

Affiliations
1 Mother Teresa Women's University, Kodaikanal, Tamil Nadu, IN
2 School of Physics, Madurai Kamaraj University, Madurai, Tamil Nadu, IN

Abstract

Document analysis and optical character recognition (OCR) systems have been under research for few decades. Especially for distorted characters, still it remains a highly challenging task to implement an OCR that works under all possible conditions and gives highly accurate results. In this paper, we investigate the application of hybrid technique that recognizes distorted broken Tamil characters. Entire application is carried out in two different phases. Parker's shape tracing along with line tracing features was included as first phase. Neural Network training along with selective thresholding minimum distance technique (MDT) for recognition was continued as second phase. The results show that the proposed hybrid technique can improve character recognition accuracy of distorted documents containing broken characters and feasible with other existing techniques, easy for extension, and may be very effective for non-headline based complex Indic scripts.

Keywords

Broken Character, Shape Tracing, Line Tracing, Minimum Distance Technique, Non-Headline Scripts.

Full Text

     

R. Indra Gandhi ¹, K. Iyakutti ²

Affiliations
1 Department of Computer Science, Mother Teresa Women's University, Kodaikanal, Tamil Nadu, IN
2 School of Physics, Madurai Kamaraj University, Madurai, Tamil Nadu, IN

Abstract

Segmentation is an important phase towards the designing of optical character recognition system. One of the important reasons for poor recognition rate in OCR system is due to incorrect segmentation of characters. Most of the segmentation algorithms primarily aim at segmenting text, graphics, page, line and word. Character segmentation is the fundamental process in character recognition approaches, which rely on isolated characters. Sometimes during segmentation, characters of same word touch each other thus producing vertically overlapping characters. In Tamil scripts, applying the simple concept of vertical projection in segmenting the whole document into individual characters does not work well. As a first step in resolving this, this paper presents an intelligent technique for solving the key problems of distorted merging (touching) characters segmentation. The results show that the proposed algorithm yields promising segmentation output and feasible with other existing techniques, easy for extension, and may be very effective for non-headline based complex Indic scripts.

Keywords

Segmentation, Distorted Character Segmentation, Vertical Overlapping, Merging (Touching) Character Segmentation, Non-Headline Scripts.

Full Text

     

R. Indra Gandhi ¹, K. Iyakutti ²

Affiliations
1 Department of Computer Applications, G.K.M College of Engineering and Technology, Tamil Nadu, IN
2 School of Physics, Madurai Kamaraj University, Tamil Nadu, IN

Abstract

The objective of this paper is to address a special problem in character recognition from document images were the reverse side scripts appear as noise on front side and even interfere with front side characters. Due to strong background artifacts so much of double-sided distortion is noticed in ancient documents. These are often caused by the so-called bleed-through effect. Even in well-preserved documents, a similar effect called show-through is noticed because of poor paper quality. These distortions must be removed to improve readability. We propose a new NNKSOM based hybrid technique, which incorporates statistical and diffusion model to deal with bleed-through grayscale document images. The proposed method proves to perform well regardless of the intensity differences between foreground and background. This is extremely useful for researchers engaged in recognizing the distorted documents in any script worldwide as the same kind of distortion can be found in most of the scripts used in the world.

Keywords

Distortion, Bleed-Through, Show-Through, Statistical and Diffusion Model.

Full Text

     

R. Indra Gandhi ¹, K. P. Kaliyamurthie ¹

Affiliations
1 Department of Computer Science and Engineering, Bharath University, Chennai, Tamil Nadu, IN

Abstract

In the recent trend steganography plays an important role in providing secrecy in internet. Steganography works by replacing bits of useless or unused data in regular computer files with bits of different, invisible information. Cryptography is used to scramble the information sent and make it as a scrambled text which is not understandable. Steganography is used to hide the information so that secrecy can be established. If Steganography and Cryptography are used together it has more advantage for providing security for the information. In this paper hiding of data in digital images formation carried out using adaptive hiding with the optimum pixel adjustment (OPA) algorithm. Resultant system evidences remarkable hiding rates with other Steganographic systems.

Keywords

Data Hiding, IWT, OPA, PSNR, Secret Key, Steganography.

Full Text

   

R. Indra Gandhi ¹, K. Iyakutti ²

Affiliations
1 Department of Computer Science, Mother Teresa Women’s University, Kodaikanal-624102, TN, IN
2 School of Physics, Madurai Kamaraj University, Madurai-62502, TN, IN

Abstract

This paper presents a new approach of Kohonen neural network based Self Organizing Map (SOM) algorithm for Tamil Character Recognition. Which provides much higher performance than the traditional neural network. Approaches: Step 1:It describes how a system is used to recognize a hand written Tamil characters using a classification approach. The aim of the pre-classification is to reduce the number of possible candidates of unknown character, to a subset of the total character set. This is otherwise known as cluster, so the algorithm will try to group similar characters together. Step 2:Members of pre-classified group are further analyzed using a statistical classifier for final recognition. A recognition rate of around 79.9% was achieved for the first choice and more than 98.5% for the top three choices. The result shows that the proposed Kohonen SOM algorithm yields promising output and feasible with other existing techniques.

Keywords

Handwritten Character, SOM, Baseline, Statistical, Structural, Crux, Meticulous and Sobel Edge Detection.

Full Text