Abstract
Braille recognition is the ability to recognize Braille cells in Braille images. The result will be used in several applications such as Braille translating, Braille copying...etc. However, the performance of these applications will be affected by some factors such as skew of Braille documents through the scanning or embossing. In this work, we designed a new approach for adjusting Braille image skewness. This approach is summarized in three steps: (1) A Braille image is segmented by estimated thresholds values which are resulting from Beta distribution, (2) simple Braille cells are detected by using morphological operations: and (3) the rotation angle of Braille image is estimated by calculating the median of slope angles of all straight lines that pass through any three dots of detected Braille cells. The proposed method is tested on several Braille images. Results demonstrated that the method is able to adjust the skewed Braille images accurately.