Guillen Torres’ Post

View profile for Guillen Torres, graphic

Researcher - OSINT against Disinformation, Media Studies Department, University of Amsterdam.

Edit: code with a different model: https://lnkd.in/etymQt_t Open Source Investigators have developped a few methods to reduce their (or their audience's) exposure to distressing material. For example, Bellingcat usually alters the photos they include in their articles (e.g. https://lnkd.in/erRVfJqM, thanks to Sarah in bellingcat's discord for the heads up). While there are tools available to automatically process images, removing or blurring bodies so that you don't have to see them many times while trying to pick up other visual clues, I couldn't find a tool to process videos. Here's a first attempt at one, making use of maskrcnn. It includes what I hope is a handy guide for the non-coders (like me!). https://lnkd.in/eepmWjzE The use case: you need to watch a distressing video more than once, but would like to not see bodies. It works reasonably well, though I would love to hear from anyone willing to help me improve it!

  • No alternative text description for this image
Matthew Lawrence

Programme Director driving OSINT for human rights accountability || Consultant helping tech and data companies build intelligence tradecraft || PhD Candidate studying OSINT’s place in the world.

1w

Something we do at Centre for Information Resilience is put things into greyscale too. It really helps with the disconnection necessary when working with this kind of content daily. Some of our other policies on the subject can be found here: https://www.info-res.org/_files/ugd/186c16_04549636a98d40a8a6231e06a15d2872.pdf

Charlotte M.

Working towards the media of the future. Love discussing #Representation #RealVoices #NewTechnology #SocialImpact #MediaLiteracy

1w

This is great and really needed! Would you mind if I share this with the Bellingcat community if you or Sarah haven't already? A small privacy question, when using the interface where are the videos stored as I note you have to upload. Or is the gdrive link up where the files remain? Are the files deleted from the system memory after a certain point like a lot of file conversion software? May be worth adding a note on this as with our autoarchiver for example (and yes the purpose of this is to save) we made sure people knew that Bellingcat staff cannot see the files people are archiving despite building the tool.

Like
Reply

Low tech Obfuscation: stick of post it note over the offending area, alternatively if using VLC set spatial blur to max and turn black and white (grey scale) by pressing CTRL + Windows key + C

Guillen Torres - let’s connect, we’re building media technology for OSINT investigators!

Like
Reply
Nathaniel F.

OSINT is fun. ˿̴̵̶̷̸̡̢̧̨̛̖̗̘̙̜̝̞̟̠̣̤̥̦̩̪̫̬̭̮̯̰̱̲̳̹̺̻̼͇͈͉͍͎̀́̂̃̄̅̆̇̈̉̊̋̌̍̎̏̐̑̒̓̔̽̾̿̀́͂̓̈́͆͊͋͌̕̚ͅ͏͓͔͕͖͙͚͐͑͒͗͛ͣͤͥ͘͜͟͢͝͞͠͡

1w

Peter Tyson

See more comments

To view or add a comment, sign in

Explore topics