README.md

metadata

title: 🕵️ Jerry - Deepfake Detection Tool
emoji: 🔍
colorFrom: blue
colorTo: purple
sdk: gradio
sdk_version: 5.49.1
app_file: app.py
pinned: false

🕵️ Jerry - Deepfake Detection Tool

A powerful web application that helps you detect whether facial images are real or AI-generated using advanced deep learning technology.

🔍 Features

• Fast Analysis: Quick detection with real-time results
• User-Friendly Interface: Simple file upload system with intuitive design
• Confidence Scoring: Detailed probability scores for both classifications with visual progress bars
• Clear Verdict: Color-coded results (green for real, red for deepfake) with confidence percentages
• Automatic Model Loading: Optimized performance with cached models using Streamlit's cache_resource
• Responsive Design: Works seamlessly on different screen sizes
• Modern UI: Clean, professional interface with custom styling

🚀 How to Use

1. Upload Image: Click "Browse files" to upload a facial image (JPG, PNG, JPEG, BMP)
2. Analyze: Click the "🔍 Analyze Image" button
3. View Results: See confidence scores and final verdict in the results panel

📊 Understanding Results

• ✅ REAL IMAGE: The image is likely authentic (displayed in green)
• 🚨 DEEPFAKE DETECTED: The image appears to be AI-generated (displayed in red)
• Confidence Scores: Percentage indicating how confident the model is in each classification
• Visual Progress Bars: Easy-to-understand graphical representation of confidence levels

🛠️ Installation

To run locally:

# Clone the repository
git clone <repository-url>
cd jerry-deepfake-detector

# Install dependencies
pip install streamlit transformers torch pillow

# Run the application
streamlit run app.py