hmnshudhmn24/real-fake-ai-generated-art-images

🎨 Real and Fake (AI-Generated) Art Images Dataset

21,642 balanced images — 10,821 real artworks and 10,821 AI-generated images — for training models to distinguish authentic art from GAN-generated fakes.

🧭 Overview

This dataset is part of the FauxFinder project, designed to build advanced models capable of distinguishing between authentic artworks and AI-generated images. Ideal for binary classification, GAN research, and computer vision benchmarking.

🗂️ Classes

✅ Key Features

• —Perfectly balanced classes (10,821 each)
• —All images resized to 256×256 pixels
• —Ready for CNN and deep learning pipelines
• —Diverse art styles, genres, and periods
• —Diverse GAN architectures for fake images

📁 Dataset Structure

Data/ ├── REAL/ # 10,821 authentic artwork images │ ├── image1.jpg │ ├── image2.jpg │ └── ... ├── FAKE/ # 10,821 AI-generated images │ ├── image1.jpg │ ├── image2.jpg │ └── ...

⚡ Quick Start

import os import pandas as pd from PIL import Image import matplotlib.pyplot as plt

Build dataset dataframe

data = [] for label in ["REAL", "FAKE"]: for file in os.listdir(f"Data/{label}"): data.append({"filename": file, "label": label})

df = pd.DataFrame(data) print(df["label"].value_counts())

Display sample image

img = Image.open("Data/REAL/image1.jpg") plt.imshow(img) plt.title("Real Artwork") plt.axis("off") plt.show()

🔧 Recommended Setup

from sklearn.modelselection import traintest_split

80-20 train-test split

traindf, testdf = traintestsplit( df, testsize=0.2, randomstate=42, stratify=df["label"] ) print(f"Train: {len(traindf)} | Test: {len(testdf)}")

💡 Use Cases

• —Real vs AI-generated art classification
• —GAN output detection research
• —Binary image classification practice
• —Transfer learning with ResNet, EfficientNet, ViT
• —Deepfake detection in artistic domain
• —Benchmarking CNN architectures