International Journal of Innovative Technology and Exploring Engineering (IJITEE) ISSN: 2278-3075, Volume-3, Issue-1, June 2013 5 Abstract— This paper presents a novel Bangla Character Recognition (BCR) system which converts a Braille Document into Bengali text which is not attempted in research work so far. The system is capable of doing the extraction of Braille Characters from a Braille document followed by decoding them into Bengali characters and then the decoded Bengali characters are normalized to Bengali text which is in human-understandable form. This system can be very useful for the blind communities and the associated persons who want to know the Braille system through Bengali language. The proposed methodology has been tested on the Braille documents collected from the Jhunka Pratibondhi Aloke Niketan, West Bengal. Index Terms—Braille Cell, Braille Notation and Bengali Character, Decoding, Normalization, Bengali Text. I. INTRODUCTION The Braille system, derived in 1821 by Frenchman Louis Braille, is a method that is widely used by visually impaired people. Braille refers to an approach in which text is printed on a thick sheet of paper using special symbols representing the letters of the alphabet [1]. A Braille Cell is composed of 6 dots arranged in three rows where each row consists of 2 columns [1][2]. However in the early days, Braille used 8 dots to represent a character. Later it was reduced to 6 because a person could read only 6 dots comfortably in one touch. Fig. 1 shows such a Braille cell with 6 dots. Figure 1. A Braille cell with 6 dots Using 6 dots, 64 different characters of a language could be represented using Braille. Braille characters are embossed on a sheet and are read by sensing the embossment by fingers. The thickness of the sheet is used for embossing the Braille is directly is proportional to the life of the document. Manuscript received on June, 2013. Dr. Santanu Halder, Mr. Abul Hasnat, Assistant Professor, Dept. of Computer Science and Engineering, Government College of Engineering Textile Technology, Berhampore, West Bengal, India. Ms. Amina Khatun, Student, M. Tech, Jadavpur University, West Bengal, India. Dr. Debotosh Bhattacharjee, Associate Professor, Dept. of Computer Science and Engineering, Jadavpur University, Kolkata, India. Dr. M. Nasipuri, Professor, Dept. of Computer Science and Engineering, Jadavpur University, Kolkata, India. Less the thickness, less the number of times the document could be read. Even though every language has its own Braille notation, very little research work has been observed in the literature survey [3]. The available optical Braille character recognition is focused only for English Braille, not for any other language [3] and hence most of the visually impaired people in the developing country are deprived of the modern technology that could be helpful for the advancement of their educational infrastructure and consequent joining the mainstream life through employment opportunities. Moreover people who work with the visually impaired person and are unable to read the Braille character require conversion of Braille documents into a normal language representation. Bengali is such a language for which no Braille Character recognition system has been developed so far. To give a solution in this direction this paper aims to design a Braille Character Recognition system (BCR) from Braille notation which will be useful for the blind community who wants to see the world using Bengali language. The BCR system comprises of three steps: (i) Braille Character Extraction from a Braille Document (ii) Decoding of extracted Braille Characters for Bengali Character Recognition (iii) Normalization of the decoded Bengali characters for getting the Bengali text. This paper is organized as follows: The section II presents the brief description of the related work. Section III presents dimension of a Braille cell. Section IV depicts the Braille notation and Bengali character. Section V describes The Bangla Character Recognition (BCR) System. Section VI depicts the Extraction phase. The decoding phase is shown in Section VII. Section VIII describes the Normalization phase. The experimental result is shown in Section IX and finally section X concludes and remarks about some of the aspects analyzed in this paper. II. RELATED WORK In 1988, Dobus and his team designed an algorithm called Lectobraille which translates relief Braille into an equivalent printed version on paper [3]. In 1993 & 1994, Mennens and his team designed an optical recognition system which recognized Braille writing. It used scanner for digitizing the Braille document [5-6]. In 1999 Ng and his team approached the problem using boundary detection techniques to translate Braille into English or Chinese [7]. The paper presented an automatic system to recognize the Braille pages and convert the Braille documents into English/Chinese text for editing. They separated embossing on two sides of a sheet into two standard templates. The system took the advantage of regular spacing between Braille dots within a cell, and the regular spacing between cells. The paper did not discuss anything about grid deformation. The capturing device used in this experiment was a digital Development of a Bangla Character Recognition (BCR) System for Generation of Bengali Text from Braille Notation Santanu Halder, Abul Hasnat, Amina Khatun, Debotosh Bhattacharjee, Mita Nasipuri 1 2 3 4 5 6