Speech Recognition Projects - Naan Mudhalvan

About Naan Mudhalvan

Naan Mudhalvan is a pioneering initiative by the Government of Tamil Nadu, India, aimed at enhancing the employability of students through industry-aligned training programs. The program collaborates with leading technology companies and educational institutions to provide cutting-edge training in various domains, including Artificial Intelligence, Machine Learning, and Speech Recognition.

About GUVI

GUVI (Grab Your Vernacular Imprint) is a leading ed-tech platform that provides vernacular language-based technical education. As a partner in the Naan Mudhalvan program, GUVI offers specialized courses in emerging technologies, including their comprehensive Speech Recognition course that covers both theoretical concepts and practical implementations.

Project Overview

This repository contains three major projects developed as part of the Naan Mudhalvan Speech Recognition course:

1. Audio Data Preprocessing and Augmentation (nm-unit1/)

   • Comprehensive audio data preparation pipeline
   • Format conversion and standardization
   • Data augmentation techniques
   • Metadata management

2. Accent-Aware Speech Recognition (nm-unit4/)

   • Advanced ASR system with accent adaptation
   • Deep learning-based implementation
   • Support for multiple English accents
   • Performance evaluation and metrics

3. Call Clarity Monitor (nm-unit5/)

   • Real-time call quality monitoring system
   • Speech recognition and analysis
   • Content safety monitoring
   • Text analytics integration
   • Django-based backend implementation

Project Structure

nm-projects-speech-recognition/
├── nm-unit1/                    # Audio Preprocessing Project
│   ├── Speech-to-Text Transcription System.ipynb
│   ├── README.md
│   ├── LICENSE
│   ├── CONTRIBUTING.md
│   └── requirements.txt
│
├── nm-unit4/                    # Accent-Aware ASR Project
│   ├── Accent-Aware Speech Recognition.ipynb
│   ├── README.md
│   ├── LICENSE
│   ├── CONTRIBUTING.md
│   └── requirements.txt
│
├── nm-unit5/                    # Call Clarity Monitor
│   ├── call-clarity-monitor/    # Frontend application
│   ├── call_clarity_backend/    # Django backend
│   ├── analyzer/               # Analysis modules
│   ├── media/                 # Media storage
│   ├── documentation.md       # Detailed documentation
│   ├── about.md              # Project information
│   ├── requirements.txt      # Dependencies
│   └── manage.py            # Django management script
│
└── README.md                    # This file

Course Coverage

The Naan Mudhalvan Speech Recognition course, in collaboration with GUVI, covers:

1. Fundamentals of Speech Processing

   • Audio signal processing
   • Feature extraction
   • Data preprocessing techniques

2. Machine Learning for Speech

   • Deep learning architectures
   • Model training and optimization
   • Performance evaluation

3. Practical Implementation

   • Real-world applications
   • Industry best practices
   • Project-based learning
   • Full-stack development
   • System integration

Getting Started

1. Clone the repository:

   git clone https://github.com/itzi-vignesh/nm-projects-speech-recognition.git
   cd nm-projects-speech-recognition

2. Install dependencies for each project:

   # For Audio Preprocessing Project
   cd nm-unit1
   pip install -r requirements.txt
   
   # For Accent-Aware ASR Project
   cd ../nm-unit4
   pip install -r requirements.txt
   
   # For Call Clarity Monitor
   cd ../nm-unit5
   pip install -r requirements.txt

3. Follow the individual project READMEs for detailed setup and usage instructions.

Contributing

We welcome contributions! Please see the CONTRIBUTING.md file for guidelines on how to contribute to this project.

License

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgments

• Naan Mudhalvan Program for providing the learning opportunity
• GUVI for their comprehensive course content and support
• Common Voice dataset for providing the training data
• Open-source community for various tools and libraries
• Django framework and its contributors
• OpenAI and other API providers

Contact

For questions and feedback, please open an issue in the GitHub repository or contact:

• GitHub: itzi-vignesh
• Project Repository: nm-projects-speech-recognition