The HFD-8000

The HFD-8000 dataset consists of 8000 labeled human face images aimed at addressing the challenges in detecting image forgeries, particularly in distinguishing between real and GAN-generated images. The dataset is structured for binary classification tasks with 6400 images for training, 1600 for testing, and a description file containing the label information. All images are resized to 224x224 pixels, providing a standardized format as JPEG for training deep learning models. This dataset is a valuable resource for advancing research in detecting deepfakes and AI-generated image forgeries.