Dot Detection of Optical Braille Images for Braille Cells Recognition

Abstract

Braille is a tactile format of written communication for people with low vision and blindness worldwide. Optical Braille Recognition (OBR) offers many benefits to Braille users and people who work with them. This paper presents a new algorithm for detecting dots composing Braille characters in an image of embossed Braille material obtained by an optical scanner. We assume that a mixture of Beta distributions can model the histogram of a scanned Braille document. The core of the proposed method is the use of stability of thresholding with Beta distribution to initiate the process of thresholds estimation. Segmented Braille image is then used to form a grid that contains recto dots and another one that contains verso dots. Using segmented image, Braille dots composing characters on both single-sided and double-sided documents are automatically identified from those grids with excellent accuracy.