Open-source model for AI vs Real Image Detection - Free deployment for accurate detection of AI-forged images

Ai Vs Real Image Detection

Developed by dima806

An image classification model based on Vision Transformer architecture, specifically designed to detect AI-generated fake images

Image Classification

Transformers

Open Source License:Apache-2.0 #Deepfake Detection #High-Precision Classification #Image Authenticity Verification

Downloads 931

Release Time : 10/15/2023

Model Overview

This model utilizes the ViT architecture to distinguish between real images and AI-generated fake images by analyzing image features. It is suitable for content moderation and security verification scenarios.

Model Features

High-Precision Detection

Achieves 98.25% accuracy and 0.982 F1 score on test datasets

Concept Drift Warning

Provides explicit warnings about potential technical drift issues with modern AI-generated images

Adjustable Threshold

Recommends adjusting classification thresholds to adapt to new types of fake images

Model Capabilities

Image Authenticity Analysis

Deepfake Detection

Binary Image Recognition

Use Cases

Content Security

Social Media Content Moderation

Automatically identifies AI-generated fake images on platforms

Filters over 98% of fake images (based on test data)

Digital Forensics

News Image Verification

Verifies whether news images are AI-generated

Provides credibility scores and classification results

🚀 Image Authenticity Detection Model

This project provides a model to check whether an image is real or AI - generated.

🚀 Quick Start

The model checks whether the image is real or fake (AI - generated).

✨ Features

Accurate Classification: The model can accurately classify real and AI - generated images, with high precision and recall rates.
Model Base: It is based on the google/vit - base - patch16 - 224 - in21k model.

📦 Installation

No specific installation steps are provided in the original document, so this section is skipped.

💻 Usage Examples

No code examples are provided in the original document, so this section is skipped.

📚 Documentation

Model Metrics

Property	Details
Model Type	Image authenticity detection model
Training Data	Not specified, but note that the data was collected about 2 years ago
Metrics	Accuracy, F1

Classification Report

Classification report:

              precision    recall  f1 - score   support

        REAL     0.9868    0.9780    0.9824     24000
        FAKE     0.9782    0.9870    0.9826     24000

    accuracy                         0.9825     48000
   macro avg     0.9825    0.9825    0.9825     48000
weighted avg     0.9825    0.9825    0.9825     48000

Note for Production Use

⚠️ Important Note

This model is trained on a dataset collected about 2 years ago. Since then, there has been remarkable progress in generating deepfake images with common AI tools, resulting in a significant concept drift. To mitigate that, you are urged to retrain the model using the latest available labeled data. As a quick - fix approach, simply reducing the threshold (say from default 0.5 to 0.1 or even 0.01) of labelling an image as a fake may suffice. However, you will do that at your own risk, and retraining the model is the better way of handling the concept drift.

💡 Usage Tip

For more details, please refer to [this Kaggle notebook](https://www.kaggle.com/code/dima806/cifake - ai - generated - image - detection - vit).

Model Visualization

![image/png](https://cdn - uploads.huggingface.co/production/uploads/6449300e3adf50d864095b90/bbtmz7duMA6o4HfEp_vjz.png)

📄 License

This project is licensed under the Apache 2.0 license.

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご