DeepFake-o-meter

An open platform integrating state-of-the-art DeepFake detection methods



If you are a developer and want your DeepFake detection method to be included in DeepFake-o-meter, please follow these steps.

Please find instructions for using DeepFake-o-meter in this video tutorial.

Upload Video

You can enter a YouTube video URL or upload a video from your computer. Note that the video should contain only one subject and you should upload only one video each time.

•Upload an url.
  Video Url:

• Or upload the video (maximum size is 50MB).


Select DeepFake Detection Methods

Select the deepfake Methods.Instruction Reference

Select Methods
    Video-based methods:
  • DSP-FWA
  • FWA
  • VA
  • Selim
  • WM
    • Image-based methods:
  • XceptionNet
  • CNNDetection
    • Both:
  • Capsule
  • ClassNSeg
  • MesoNet
  • Upconv




Input your e-mail and PIN code

The submission notification and a download link for your results will be sent to your email. Only one email is required.

• E-mail address:

The PIN code is used for download verification

• A 4~6 digits PIN code:


Agreement

By clicking this box, I accept the terms of using DeepFake-o-meter.

Input email address and PIN code to check the status of your submission.




DeepFake-o-meter is developed by Yuezun Li, Cong Zhang, Pu Sun, Lipeng Ke, Yan Ju, and Siwei Lyu of UB Media Forensic Lab. If you are interested in the design and implementation of DeepFake-o-meter, please check out our paper [Link]. The DeepFake-o-meter platform is grateful for the tremendous technical support of the CSE-consult group of the CSE Department at UB. Last but not least, the DeepFake-o-meter team would like to thank all the authors, contributors, and developers of the individual detection algorithms (see the full list Reference)