✋ SignifyAI: Real-Time American Sign Language Recognition

📌 Overview

SignifyAI is an intelligent system that recognizes American Sign Language (ASL) hand gestures using deep learning and computer vision.
It allows users to upload ASL gesture images or use their webcam for real-time recognition, translating hand gestures into corresponding English alphabets.

---

🌟 Features

✅ ASL Alphabet Recognition (A–Z + special signs)
✅ Upload Image Prediction – Detects gestures from uploaded images
✅ Live Webcam Mode – Real-time ASL recognition directly from camera
✅ Interactive Web Interface built using Flask
✅ Pretrained ResNet18 model fine-tuned for ASL dataset
✅ Data Augmentation for better lighting & background adaptability
✅ Automatic Cleanup of uploaded files after prediction

🏗️ Project Structure

SignifyAI: Real-Time American Sign Language Recognition/
│
├── model/
│   └── asl_model_files_re/
│       ├── asl_model.pth
│       └── asl_classes.json
|
|__ notebook/
|   |__ asl-signlanguage-detector.ipynb
│
├── static/
│   ├── css/
│   ├── js/
|   |   |___quiz.js
|   |   |___script.js
│   └── images/
│
├── templates/
│   ├── index.html
│   ├── upload.html
│   ├── upload_result.html
│   ├── recognize.html
│   ├── alphabets.html
│   └── quiz.html
│
├── uploads/               # temporary folder for uploaded files
│
├── app.py                 # main Flask app
├── requirements.txt       # dependencies
└── README.md

---

🧩 Model Details

• Architecture: ResNet18

• Framework: PyTorch

• Training Augmentations:

  • Random rotations
  • Color jitter
  • Random affine transformations
  • Normalization with ImageNet stats

• Validation Transform: Resize → Normalize → Tensor

• Accuracy: High accuracy on clean dataset and improved real-world performance after retraining.

---

🎬 Demo Videos

📸 1. Website Walkthrough:
👉 https://github.com/user-attachments/assets/73fef755-8e1f-4bc5-b384-34790c39c9a2

🖼️ 2. Upload Recognition (Image Prediction):
👉 https://github.com/user-attachments/assets/2099a1af-6b1d-453a-870c-20049fd8fc22

📹 3. Live Webcam Recognition:
👉 https://github.com/user-attachments/assets/981924ae-1a94-4d59-a4e7-f2bc925a8ddb

---

⚙️ How to Run

1️⃣ Clone the repository

git clone https://github.com/JaweriaAsif745/-SignifyAI-Real-Time-American-Sign-Language-Recognition.git
cd ASL-Recognition-System

2️⃣ Install dependencies

pip install -r requirements.txt

3️⃣ Run the Flask app

python app.py

4️⃣ Open in browser

Go to: http://127.0.0.1:5000/

---

🧩 Usage 🔹 Image Upload Mode

Open the Upload page. Upload a hand gesture image. The model predicts the alphabet and confidence score.

🔹 Live Webcam Mode

Go to the Recognize page. Allow camera permissions. Show your ASL gesture in front of the webcam. See live predictions appear instantly.

🖼️📸 Screenshots

🏠 Home Page

🧠 Alphabets Page

📹 Quiz Page

📤 Image Upload Page

📤 Image Result Page

🤖 Webcam Detection

---

🧰 Tech Stack

Category	Tools
Backend	Flask, Python
Frontend	HTML, CSS, JavaScript
AI/ML	PyTorch, ResNet18
Deployment	Localhost (Flask)
Dataset	ASL Alphabet Dataset

---

🧾 License

This project is licensed under the Creative Commons Attribution-NonCommercial 4.0 International License (CC BY-NC 4.0).

💡 Future Improvements

• Implement word-level ASL translation (sequence-based using video frames).
• Add voice output (text-to-speech).
• Improve real-time performance using ONNX or TensorRT.

---

👩‍💻 Author

Jaweria Asif 📍 Pakistan 💬 “Building AI that understands human gestures.” 🔗 https://github.com/JaweriaAsif745

---

⭐ If you like this project, give it a star on GitHub!

---