Abstract
With the advent of deepfakes videos, it is important for the purposes of promoting societal trust to be able to verify the authenticity of a person who appears in a social media audiovisual stream or a broadcast/cable television video. This disclosure describes techniques that enable the viewer of a particular piece of video purporting to show a person to verify that the video indeed represents the person authentically in all respects, e.g., that the person truly made the statements heard in the video and did so with their authentic tone, cadence, delivery, body language, etc. A video camera certified as being owned and operated by the person in the audiovisual stream records the video, digitally signs it, and publishes it to a blockchain ledger. A signed beacon, such as a changing QR code displayed during the recording, is captured by other video cameras in the room. A video is verified as genuine by verifying the signature of the beacon and by semantically matching the recording with the video stored in the blockchain.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
Belmon, Stephane and Rammohan, Roshan, "Deepfake-Resistant Video Using Blockchain Ledgers and Machine Learning", Technical Disclosure Commons, (April 11, 2024)
https://www.tdcommons.org/dpubs_series/6869