# Bidirectional-Indian-Sign-Language-Translator

**Repository Path**: zbgit11/Bidirectional-Indian-Sign-Language-Translator

## Basic Information

- **Project Name**: Bidirectional-Indian-Sign-Language-Translator
- **Description**: No description available
- **Primary Language**: Unknown
- **License**: Not specified
- **Default Branch**: main
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2026-10-08
- **Last Updated**: 2026-10-08

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# Bidirectional Indian Sign Language Translator

This repository contains the source code for a comprehensive Bidirectional Indian Sign Language (ISL) Translator. The project aims to bridge the communication gap for the deaf and mute community by enabling seamless translation between text and Indian Sign Language.

🏆 Submitted as part of National Smart India Hackathon (SIH) 2024 Finals:

### Use Case Diagram
![Use Case Diagram](Images/usecase_diagram.png)  

The included prototype use case diagram:
docs/Prototype_Usecase_Diagram.png

## Project Overview

The system consists of two main modules:

1.  **Text to Indian Sign Language (Text-to-ISL)**:
    -   Converts input text (English) into ISL animations performed by a digital avatar.
    -   Utilizes Natural Language Processing (NLP) to restructure sentences according to ISL grammar (Subject-Object-Verb).
    -   Maps words to corresponding SiGML (Signing Gesture Markup Language) files which drive the avatar's actions.

2.  **Indian Sign Language to Text (ISL-to-Text)**:
    -   Recognizes hand gestures from a webcam feed and translates them into readable text.
    -   Leverages MediaPipe for hand tracking and custom-trained TensorFlow models for gesture classification.
    -   Provides real-time feedback and translation.

## Repository Structure

The codebase is organized as follows:

-   **[Indian-Sign-Language-to-Text/](./Indian-Sign-Language-to-Text/)**: Contains the Flask application and machine learning models for gesture recognition.
    -   Includes `app.py`, `model/` (TFLite classifiers), and `Notebooks/` for training.

-   **[Text-to-Indian-Sign-Language/](./Text-to-Indian-Sign-Language/)**: Contains the subsystem for generating ISL animations from text.
    -   **[Text-to-ISL-NLP/](./Text-to-Indian-Sign-Language/Text-to-ISL-NLP/)**: A Python/FastAPI backend for NLP processing and grammar conversion.
    -   **[Web-UI/](./Text-to-Indian-Sign-Language/Web-UI/)**: A Node.js/Express frontend interface for user interaction and avatar display.

## Getting Started

To run the complete system, you will need to set up and run the individual components separately. Please refer to the `README.md` files in each subdirectory for detailed setup and usage instructions.

### Prerequisites

-   **Python 3.8+**: Required for the NLP backend and ISL-to-Text application.
-   **Node.js & npm**: Required for the Text-to-ISL web interface.
-   **Webcam**: Required for the ISL-to-Text gesture recognition feature.

### Quick Links

-   [Setup Text-to-ISL System](./Text-to-Indian-Sign-Language/README.md)
-   [Setup ISL-to-Text System](./Indian-Sign-Language-to-Text/README.md)

## Technologies Used

-   **Frontend**: HTML, CSS, JavaScript, Node.js, Express.
-   **Backend**: Python, Flask, FastAPI.
-   **Machine Learning / AI**: TensorFlow, MediaPipe, scikit-learn, spaCy, NLTK.
-   **Translation**: Google Translate API (`googletrans`).

## About

This project focuses on accessibility and inclusion, providing a bidirectional communication tool for Indian Sign Language users. It combines computer vision for sign recognition and rule-based NLP for sign generation to create a robust translation system.