Fooling Computer Vision
Published January 22, 2020
|
25 min
    Download
    Add to queue
    Copy URL
    Show notes

    Wiebe van Ranst joins us to talk about a project in which specially designed printed images can fool a computer vision system, preventing it from identifying a person.  Their attack targets the popular YOLO2 pre-trained image recognition model, and thus, is likely to be widely applicable.

      15
      15
        0:00:00 / 0:00:00