Image Pre-Processing Effect on OCR's Performance for Image Conversion to Braille Unicode

Abstract

The number of cases of visual impairment has been rising, particularly in Indonesia. Additionally, most disabilities in Indonesia are visual impairments, according to data from the Population Census. Blindness is one aspect of this condition that accounts for a sizable percentage of it. Considering this, we are making an effort to assist them by translating literature, which entails converting text included in images into braille Unicode. A braille embosser, a device that can print braille sheets for the blind to read, will then be used to print this translation into braille. In this study, we discuss the consequences of pre-processing while converting a text-containing image to braille Unicode. Here, we employ the pre-processing techniques of binarization, denoising, and skew correction. These techniques are used before using the OCR, and the results are documented in this study. Additionally, the outcomes of pre-processed and raw photos are contrasted to determine the efficacy of applying pre-processing to an image.