vtl icon indicating copy to clipboard operation
vtl copied to clipboard

Fake videos detection by tracing the source using video hashing retrieval.

ViTHash

📜 Directory

  • Introduction
  • ViTHash
  • Train or Test
    • Train
    • Test
  • Downloads
    • Download Models and Datasets
    • DeepFake Dataset Samples
    • Video Inpainting Dataset Samples
    • Video Splicing Dataset Samples

✨ Introduction

An effective and reliable method for forgery video detection.

👀️ViTHash

🔬 Train or Test

Train

Train ViTHash

python train_h.py --local_rank=0 --path=../Celeb-DF --type=Celeb-DF --hash_bits=512 --pretrained=x

Test

Test ViTHash

python test_h.py --local_rank=0 --path=../Celeb-DF --type=Celeb-DF --hash_bits=512 --pretrained=x

🚀️ Downloads

Download Models and Datasets

Pretrained models, hash centers and datasets

Item Link Comments
All Models BaiduNetdisk code:i5l1
Video Splicing BaiduNetdisk code: a9in
DeepFake BaiduNetdisk code: eaj0
Video Inpainting BaiduNetdisk code: ande Extension ofDAVIS2016

DeepFake Dataset Samples

The person with different scenes.

Source Videos

Fake Videos

Different fake videos are made from the same source.

Source Video

Fake Videos

Video Inpainting Dataset Samples

The first gif of the boat is the original video, and the remaining five videos are generated by different object inpainting methods.

Video Splicing Dataset Samples

The bird, person, soldier and planes are all spliced by manual.

Tips

We have developed an additional tool to compare the difference between real video and fake video. Please see tool