OhMyYuwan/face-forgery-detection

Face Forgery Detection

Detect whether a face image is real or forged using state-of-the-art deep learning models.

Features

• —Single Image Detection: Upload one image and get instant results
• —Batch Detection: Process multiple images at once
• —Multiple Models: Choose from 7 different detection models
• —Confidence Scores: See detailed prediction scores and thresholds

Models

This Space uses models from OhMyYuwan/face-forgery-detection:

• —ConvNeXt Base
• —DINOv2 Base
• —FasterViT-2
• —InceptionNeXt Base
• —InternViT-300M
• —MambaVision-T
• —MaxViT Base

Usage

1. 1.Select a model from the dropdown
2. 2.Upload an image (or multiple images for batch mode)
3. 3.Click "Detect" to see results

Citation

@article{tifs2026,
  title={Face Forgery Detection},
  year={2026}
}