Braille Detection Application Using Gabor Wavelet and Support Vector Machine

Difference any exchanging of several kinds information such visual, printed or written form between impairments vision or blind with normal will cause problem especially in written form. For instance in the assistance of a blind child by family with normal vision. One of family role is education to help their children to study and understand about their learning development from home, particularly for blind children. In this study, braille letters can be identified through images obtained using a scanner to help parents and families of a blind child in learning assistance by implementing the Gabor Wavelet feature extraction method. The features used are standard deviation, mean, variance, and median with theta angles of 00,300,450,600,1200,1350,1800 and wavelengths 3,6,13,28, and 58. These features will be combined and used as input as test data and training data. at the Support Vector Machine (SVM) classification stage and generates words in the alphabet. The braille letters detected in this study were small braille letters, capital braille letters, punctuation marks, and numbers. The test is carried out using a multi-class confusion matrix scenario to determine the level of accuracy, precision, and recall. Based on the results of tests carried out using 758 braille data, the accuracy value is 98.15%, the precision value is 97.66% and the recall value is 98.28%. From these results it can be concluded that the Gabor Wavelet feature extraction method and the Support Vector Machine (SVM) can be used to identify braille letters.