File size: 1,479 Bytes
6bb42a3
 
 
 
 
e64fe6d
 
6bb42a3
 
480ffd1
6bb42a3
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
---
title: Real vs Fake - AI Image Detector
emoji: 🔍
colorFrom: blue
colorTo: purple
sdk: docker
app_port: 7860
pinned: false
license: mit
short_description: AI-powered detector for identifying AI-generated images
---

# 🔍 Real vs Fake: AI Image Detector

Detect whether an image is real or AI-generated using perceptual features.

## How It Works

This detector uses **ARNIQA** (Attention-based distortion-aware No-Reference Image Quality Assessment)
to extract perceptual features from images. These features are then used by a trained classifier
to distinguish between real photographs and AI-generated images.

## Supported AI Models

The detector can identify images generated by:
- Stable Diffusion (various versions)
- Midjourney
- DALL-E
- And other popular generative models

## Usage

1. Upload an image (JPG, PNG)
2. The model will analyze perceptual features
3. Get instant results showing whether the image is real or AI-generated

## Model Details

- **Feature Extractor**: ARNIQA (from miccunifi/ARNIQA)
- **Classifier**: 2-layer MLP trained on GenImage dataset
- **Input Size**: Variable (automatically resized)
- **Inference**: CPU-based for universal accessibility

## Privacy

All processing happens in your browser session. Uploaded images are not stored.

## Research

Based on research using perceptual classifiers for detecting generative images.
The full research codebase is private.

## License

MIT License - Free for personal and commercial use.