Get our free extension to see links to code for papers anywhere online!

Chrome logo  Add to Chrome

Firefox logo Add to Firefox


VPN: Video Provenance Network for Robust Content Attribution

Sep 21, 2021
Alexander Black, Tu Bui, Simon Jenni, Vishy Swaminathan, John Collomosse


Share this with someone who'll enjoy it:


We present VPN - a content attribution method for recovering provenance information from videos shared online. Platforms, and users, often transform video into different quality, codecs, sizes, shapes, etc. or slightly edit its content such as adding text or emoji, as they are redistributed online. We learn a robust search embedding for matching such video, invariant to these transformations, using full-length or truncated video queries. Once matched against a trusted database of video clips, associated information on the provenance of the clip is presented to the user. We use an inverted index to match temporal chunks of video using late-fusion to combine both visual and audio features. In both cases, features are extracted via a deep neural network trained using contrastive learning on a dataset of original and augmented video clips. We demonstrate high accuracy recall over a corpus of 100,000 videos.

* CVMP2021 camera-ready version 


   Access Paper Source



Share this with someone who'll enjoy it: