Image-Based Recognition of Braille Using Neural Networks on Mobile Devices

Braille documents are part of the collaboration with blind people. To overcome the problem of learning Braille as a sighted person, a technical solution for reading Braille would be beneficial. Thus, a mobile and easy-to-use system is needed for every day situations. Since it should be a mobile system, the environment cannot be controlled, which requires modern computer vision algorithms. Therefore, we present a mobile Optical Braille Recognition system using state-of-the-art deep learning implemented as an app and server application.