Apppublic

mani880740255/AI_Image_detection

sourceHugging Faceupdated 6mo agoView on Hugging Face

0likes

App README

VisionAI — Premium AI Image Detection Dashboard

VisionAI is a modern, high-end web application designed to detect AI-generated images using state-of-the-art deep learning. It features a stunning Glassmorphism + Soft UI dashboard and integrates with Twilio to provide real-time voice call alerts when a high-confidence AI image is detected.

✨ Features

—Modern Dashboard: Premium glassmorphism aesthetic with animated cosmic backgrounds.
—AI Detection: Uses the umm-maybe/AI-image-detector (Swin Transformer) model.
—Local Inference: Supports loading models from local disk for zero-latency, private analysis.
—Real-time Alerts: Triggers Twilio voice calls when AI confidence exceeds a configurable threshold (default: 50%).
—History & Analytics: Track session stats and recent scan history.
—Mobile Responsive: Fully optimized for all device sizes.

🛠️ Tech Stack

—Frontend: HTML5, Vanilla CSS (Glassmorphism), JavaScript (ES6+).
—Backend: Python, Flask, Flask-CORS.
—AI Engine: PyTorch, HuggingFace Transformers.
—Alerting: Twilio API (Voice).

🚀 Getting Started

1. Prerequisites

—Python 3.8+
—Twilio Account (SID, Token, and Phone Number)

2. Installation

Clone the repository:

bash

git clone https://github.com/yourusername/visionai-image-detection.git
cd visionai-image-detection

Install dependencies:

bash

pip install -r requirements.txt

3. Model Setup

Since the model weights are large, you should download pytorch_model.bin (e.g., from umm-maybe/AI-image-detector on HuggingFace) and place it in the local_model/ directory, or run the setup script:

bash

python setup_local_model.py

Note: The setup script will attempt to copy your `.bin` from Downloads and fetch necessary config JSONs.

4. Environment Configuration

Create a .env file in the root directory and add your credentials:

env

TWILIO_ACCOUNT_SID=your_sid_here
TWILIO_AUTH_TOKEN=your_token_here
TWILIO_FROM_NUMBER=your_twilio_number
YOUR_PHONE_NUMBER=your_target_phone_number
AI_CONFIDENCE_THRESHOLD=0.5

5. Run the Application

Start the Flask server:

bash

python app.py

Access the dashboard at http://127.0.0.1:5000.

📂 Project Structure

—app.py: Flask backend with inference and Twilio logic.
—local_model/: Directory containing model weights and configs.
—index.html: Main dashboard UI.
—style.css: Glassmorphism design system.
—script.js: Frontend logic and API integration.
—setup_local_model.py: Script to prepare local model files.

🛡️ License

Distributed under the MIT License. See LICENSE for more information.

Created by Mani Deep