An Effective Approach of English Braille to Text Conversion for Visually Impaired Using Machine Learning Technique

EasyChair Preprint 9908

Abstract

Machine learning is a burgeoning technology where the machine learns from the historical data. Predictions are made using a variety of algorithms based on past data. Braille is a fundamental system of raised dots that enables the visually challenged and impaired (VI) to write and read by touch. In the digital age, the VI are facing great challenges in accessing digital content, integrating speech ware as output will assist them in gaining more knowledge. The existing papers, uses SVM classifier, Braille script, keypad and microphone as input and the relevant English alphabets are produced as output. This paper proposes a novel approach to convert English Grade 2 braille to English text using various different pre-processing steps using SVM Classifier. Braille images are captured by digital camera and pre-processed. The features of the scanned images are extracted by Segmentation. Support Vector Machine (SVM) Classifier with PCA feature extraction were used for Braille character recognition.

Keyphrases: Braille, Braille cell recognition., PCA, SVM classifier, visually impaired

@booklet{EasyChair:9908,
  author    = {K Arun Kumar and V Kanimozhi and G Shruthi},
  title     = {An Effective Approach of English Braille to Text Conversion for Visually Impaired Using Machine Learning Technique},
  howpublished = {EasyChair Preprint 9908},
  year      = {EasyChair, 2023}}