|
|
--- |
|
|
title: AI DeepFake Detector |
|
|
emoji: 🔍 |
|
|
colorFrom: blue |
|
|
colorTo: green |
|
|
sdk: gradio |
|
|
app_file: app.py |
|
|
pinned: false |
|
|
--- |
|
|
# AI DeepFake & Manipulation Detector (Web App) |
|
|
|
|
|
This project detects: |
|
|
- Whether an image is **AI-generated or real** |
|
|
- Whether it contains **manipulated regions** |
|
|
- Highlights edited areas with a **mask overlay** |
|
|
|
|
|
## Models Used |
|
|
- Deepfake Classifier: `prithivMLmods/deepfake-detector-model-v1` |
|
|
- Forgery / Manipulation Detector: `zhipeixu/fakeshield-v1-22b` |
|
|
|
|
|
## Features |
|
|
- Upload any image |
|
|
- Get classification (Real vs AI) |
|
|
- Fake probability (%) |
|
|
- Detect manipulated objects |
|
|
- View highlighted overlay |
|
|
|
|
|
## How to Use |
|
|
1. Upload your image into the interface |
|
|
2. Click **Analyze** |
|
|
3. View results: |
|
|
- Original Image |
|
|
- Classification + % confidence |
|
|
- Manipulation explanation |
|
|
- Mask overlay (if detected) |
|
|
|
|
|
## Setup (Hugging Face Spaces) |
|
|
- Create a Space |
|
|
- Add your HF token in Settings → Secrets: |
|
|
- Name: `HF_API_TOKEN` |
|
|
- Upload: |
|
|
- `app.py` |
|
|
- `requirements.txt` |
|
|
- `README.md` |
|
|
|
|
|
Your website will run automatically. |
|
|
|
