MIT Researchers Develop Improved Algorithm For Print On Demand Camouflage

Credit - Courtesy The ResearchersPhoto Credit – Courtesy The Researchers

Researchers at MIT have developed a new algorithm that analyzes photos of a scene taken from multiple perspectives and produces a camouflage covering for an object placed within that environment.

Objects hidden by the best performing algorithm took more than 3 seconds to find, which is significantly longer than the casual glance the camouflage is intended to fool.

Although camouflage systems like this aren’t new, this latest effort from the MIT research team is seemingly the most advanced variant to date. Technology like this could easily be adapted to cover up sensor placements, machine gun placements, or other pre-established structures and equipment in the field. However, I don’t see an application for its use with clothing, individual equipment or vehicles as it’s too specific.

Hat tip:


6 Responses to “MIT Researchers Develop Improved Algorithm For Print On Demand Camouflage”

  1. Chris says:

    They should talk to Guy Kramer about this they might learn something.

  2. Chris says:


  3. nigel tufnel says:

    i’m glad MIT figured out how to take pictures of stuff and print them out.

  4. Bussaca says:

    Not judging this either way… Wouldn’t this have been a more impressive video, if they walked up and picked up the camouflaged items after scanning by them.. I don’t know if it was the replay quality of youtube, but just showing a 15min video of Image Stills… isn’t impressive at all.. This is some Guy Cramer level stuff here..

    Any reason they couldn’t have walked by with a video camera recording live video, instead of images of a tree in a park by a lake.. Whats camouflaged? a box?

    I’m not gonna call shanangins quite yet.. but this was a bad release..from a PR stand point..

  5. m5 says:

    So, you guys are too dumb to read the linked artcile to find out what actually had been done (and why) at MIT?

    Hint: It’s about the algorithms. The video is a simulation, apparently showing the still-pics (with the different viewing angles) that were used as an input for the algorithm for creating the camo pattern on the virtual(!) box. A real-life object would have shadows…