Skip to content

Kinara2020/Vaani

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

2 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

🌸 Vaani — Voice Made Visible

Vaani is an accessibility-focused Android application that combines real-time speech transcription, Indian Sign Language (ISL) interaction, and AI-powered accessibility workflows to create a more inclusive communication experience.

The project leverages Generative AI concepts, speech recognition, and MediaPipe-based computer vision technologies to help users interact more effectively during conversations and accessibility-driven scenarios.

✨ Features

🎤 Real-Time Speech Transcription

  • Converts live speech into real-time text captions
  • Continuous speech recognition support
  • Responsive and lightweight Android implementation

💾 Offline Transcript Management

  • Save transcripts locally on the device
  • View latest conversations
  • Clear transcript history
  • Offline accessibility support

🔔 Notification-Based Accessibility Controls

Manage transcription sessions directly through Android notifications.

Supported controls include:

  • Stop transcription
  • Mute transcription
  • Quick accessibility access

This improves usability during multitasking and live interactions.

🤟 Indian Sign Language (ISL) Interaction

Vaani includes MediaPipe-powered hand tracking and gesture interaction support for Indian Sign Language accessibility workflows.

The application uses real-time camera processing to support gesture-based interaction and accessibility-focused communication.

✨ Teach Sign Feature

The application also includes a Teach Sign feature for interactive sign visualization and learning support.

🛠️ Technologies Used

Android Development

  • Android Studio
  • Java
  • XML

AI / Generative AI Technologies

  • Speech Recognition APIs
  • Real-Time Transcription Workflows
  • Accessibility-Focused AI Interaction

Computer Vision

  • MediaPipe
  • Hand Tracking
  • Gesture Processing

🚀 How It Works

  1. Launch the application
  2. Grant microphone and camera permissions
  3. Start live transcription
  4. View real-time captions
  5. Save and manage transcripts
  6. Use ISL interaction mode
  7. Manage sessions through notification controls

📱 Screenshots

(Add screenshots here)

Suggested screenshots:

  • Home Screen
  • Live Transcription
  • Notification Controls
  • ISL Interaction Mode
  • Teach Sign Feature

🎯 Problem Statement

Communication accessibility remains a challenge for many individuals during everyday interactions, especially for users with hearing or speech-related difficulties.

Vaani aims to improve accessibility by combining:

  • speech recognition,
  • computer vision,
  • Indian Sign Language interaction,
  • and Generative AI-powered workflows

into a practical Android application.

🧠 Challenges Faced

Some major challenges during development included:

  • handling continuous speech recognition,
  • optimizing Android background services,
  • reducing transcription latency,
  • integrating real-time gesture tracking,
  • and balancing accessibility with mobile performance.

🔥 Key Highlights

  • Accessibility-focused Android application
  • Real-time speech transcription
  • Offline transcript support
  • MediaPipe-powered gesture interaction
  • Indian Sign Language accessibility features
  • Notification-level accessibility controls
  • Generative AI-oriented interaction workflows

🌱 Future Improvements

Planned future enhancements include:

  • multilingual transcription,
  • improved ISL recognition accuracy,
  • AI-powered conversation summarization,
  • cloud synchronization,
  • and enhanced accessibility customization.

🎥 Demo Video

(https://youtu.be/0nIeDy_3_RM?si=FMGj31U9KHPo8j16)

📝 Blog https://dev.to/kinara2020/vaani-ai-making-communication-more-inclusive-with-speech-recognition-and-indian-sign-language-1i55

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors