🤟 Real-Time Sign Language Translator

Bridging Conversations Between Sign Language and Speech

Helping computers understand hand gestures, one sign at a time.

---

Why this project?

Communication should never be limited by language or accessibility barriers. This project uses Computer Vision and Machine Learning to recognize sign language gestures through a webcam, translate them into readable text, and instantly convert the prediction into speech. The goal is to create a fast, lightweight, and real-time communication assistant that makes conversations more inclusive.

---

How It Works

📷 Webcam

      │

      ▼

✋ Hand Detection

      │

      ▼

🧩 21 Hand Landmarks

      │

      ▼

📊 Feature Extraction

      │

      ▼

🧠 Random Forest Classifier

      │

      ▼

🔤 Predicted Alphabet

      │

      ▼

📝 Sentence Builder

      │

      ▼

🔊 Text-to-Speech

---

Project Preview

---

Features

👋 Gesture Recognition Real-time hand tracking 21 MediaPipe landmarks Fast gesture detection Stable predictions Webcam support	🧠 Machine Learning Random Forest classifier Confidence score prediction Feature extraction pipeline Dataset generation Model training support
💬 Translation Alphabet recognition Live text generation Sentence formation Continuous prediction Noise filtering	🔊 Accessibility Text-to-Speech Real-time voice output Lightweight execution Easy setup Cross-platform support

---

Technology Stack

Component	Technology
Language	Python
Computer Vision	OpenCV
Hand Tracking	MediaPipe
Machine Learning	Scikit-learn
Data Processing	NumPy, Pandas
Speech Engine	pyttsx3

---

Project Architecture

Video Stream
      │
      ▼
MediaPipe
Hand Detection
      │
      ▼
21 Landmark Coordinates
      │
      ▼
Feature Extraction
      │
      ▼
Random Forest Model
      │
      ▼
Alphabet Prediction
      │
      ▼
Sentence Generator
      │
      ▼
Speech Output

---

Project Structure

Real_Time_Sign_Language_Translator

├── dataset/
├── models/
├── static/
├── templates/
├── utils/
│
├── app.py
├── train_model.py
├── requirements.txt
└── README.md

---

Installation

Clone the repository

git clone https://github.com/Mahima-sys/Real-Time-Sign-Language-Translator.git

Go into the project

cd Real-Time-Sign-Language-Translator

Install the required packages

pip install -r requirements.txt

Run the application

python app.py

---

Current Capabilities

✅ Real-time gesture recognition

✅ Live alphabet prediction

✅ Sentence formation

✅ Confidence score display

✅ Text-to-Speech output

✅ Dataset collection

✅ Machine Learning inference

---

Future Roadmap

• Complete A to Z recognition accuracy improvements
• Word-level prediction
• Sentence-level translation
• Deep Learning based gesture recognition
• Transformer models
• Flask deployment
• Mobile application
• Voice command support
• Cloud inference
• Multi-language speech output

---

Why MediaPipe?

MediaPipe provides highly optimized real-time hand landmark detection capable of identifying 21 key hand points with remarkable speed and accuracy. Combined with a lightweight Random Forest classifier, the system delivers responsive predictions while remaining efficient enough to run on standard hardware.

---

Developer

Mohammad Fazil Firojkhan Malek

AI Developer

Machine Learning Enthusiast

Computer Vision Developer

Building technology that makes communication more accessible for everyone.

---

Support

If this project helped you, consider giving it a ⭐ and sharing feedback or suggestions for future improvements.

---

Every gesture deserves to be understood.