Nemeth Braille Mathematical Code Recognition using Deep Learning Techniques
P Jeba Malar, Kumudha Raimond, S. Deepa Kanmani · 2024
The Nemeth Braille Mathematical Code (NBMC) is an essential resource for individuals with visual impairments to comprehend scientific and mathematical symbols. It enables them to understand intricate mathematical equations, symbols, and declarations. Still, there are many challenges in accurately identifying NBMC. This makes it more difficult for those with vision problems to learn information and be self-driven. In this work, a deep learning (DL) based approach is proposed for recognizing NBMC. Comparisons are made between different deep learning models such as convolutional neural network (CNN), InceptionV3, DenseNet201, CNN combined with ResNet50, ResNet50, and VGG16; where CNN combined with ResNet50 model provides the best accuracy among them with a value of 91.20%. Then by using the layers of CNN model combined with ResNet50 works as backend for the web application. The NBMC are then translated into text-to-speech by the web application, to make it easily accessible to visually challenged people.