End-to-End Solution for Braille-To-Chinese

In the field of education, Teachers of Students with Visual Impairments need to read and review student’s Braille homework. Reading Braille consumes a lot of teachers' time and energy. We propose an end-to-end solution to translate Braille into Chinese. Our approach uses a deep learning method to detect and recognize Braille in images and translate them into Chinese and mathematical formulas. Finally, we reconstruct the translation results into the style of the input Braille image. Our approach is able to adapt to more complex Braille images and make the translation results easier to read and understand, significantly improving the speed of reading Braille for non-disabled.