OBSER : Bidirectional Translation Between Arabic & Braille

• Key Features

• Bidirectional Translation: Converts Arabic text to Braille and Braille to Arabic.

• Deep Learning-Based Recognition: Utilizes Convolutional Neural Networks (CNNs) for accurate character recognition.

• Custom Dataset Development: Creation and preprocessing of datasets containing Arabic letters, numbers, and symbols.

• Inclusive Design: Supports communication for both visually impaired and sighted users.

  
  

• Challenges & Solutions

One of the main challenges was the lack of high-quality Arabic–Braille datasets. This was addressed by generating custom datasets and applying preprocessing techniques such as resizing, denoising, and data augmentation to improve model accuracy and class balance.

• Technology Stack

• Programming & Frameworks: Python, TensorFlow

• Computer Vision: OpenCV

• Development Environments: Google Colab, Jupyter Notebook

• Data Handling: Dataset cleaning, augmentation, and balancing

  
  

• Outcomes

The system achieved accurate character recognition and near real-time translation between Arabic and Braille. OBSER contributes to accessibility and digital inclusion efforts, aligning with the goals of Saudi Vision 2030.