Deaf Aid is a web-based application built to help the deaf and hard-of-hearing community by combining Computer Vision and Speech Recognition.
Hearing-impaired individuals often face difficulties in perceiving verbal communication and recognizing human presence, especially in unfamiliar or public environments. Traditional hearing aids and communication tools do not always address the broader need for situational awareness—such as identifying who is speaking, what is being said, or if someone is approaching. This gap creates a communication barrier and increases dependency on others. Therefore, there is a need for a portable, intelligent module that can detect nearby people, identify visible objects, and transcribe spoken words in real-time to assist hearing-impaired individuals in maintaining independence and safety.
The aim of the project is to create a smart, real-time assistive website that bridges the communication gap for hearing-impaired individuals by integrating advanced AI technologies for visual and auditory interpretation. This system intends to improve accessibility, independence, and safety by enabling users to better understand their environment and interact with others more effectively. It is especially focused on being functional in everyday situations, whether at home, in public spaces, or while traveling.
- 👀 Real-time Person Detection using TensorFlow.js (COCO-SSD model)
- 🎤 Voice-to-Text Transcription using Web Speech API
- 💾 Backend Integration (Node.js + MongoDB) to store transcribed text
- 📱 Responsive UI built with HTML, CSS, and JavaScript
- 🔒 Secure and easy to use for accessibility enhancement
- Frontend: HTML, CSS, JavaScript
- AI/ML: TensorFlow.js (COCO-SSD)
- Speech Recognition: Web Speech API
- Backend: Node.js, Express.js
- Database: MongoDB (via Mongoose)