A two stage Braille Character segmentation approach for embossed double sided Hindi Devanagari Braille documents

Abstract

The Optical Braille Character Recognition (OBR) system is in significant need in order to preserve the Braille documents to make them available in future for the large section of visually impaired people and also to make the bi-directional communication between the sighted people and the visually impaired people feasible. The recognition and transcribing the double sided Braille document into its corresponding natural text is a challenging task. This difficulty is due to the overlapping of the front side dots (Recto) with that of the back side dots (Verso) in the Inter-point Braille document. In such cases, the usual method of template matching to distinguish recto and verso dots is not efficient. In this paper a new system for double sided Braille dot recognition is proposed, which employs a two-stage highly efficient and an adaptive technique to differentiate the recto and verso dots from an inter-point Braille using the projection profile method. In this paper we present (i) a horizontal projection profile for Braille line segmentation, (ii) vertical projection profile for Braille word segmentation and (iii) Integration of horizontal and vertical projection profiles along with distance thresholding for Braille character segmentation. We demonstrate the effectiveness of this segmentation technique on a large dataset consisting of 754 words from Hindi Devanagari Braille documents with varying image resolution and with different word patterns. A recognition rate of 96.9% has been achieved.