Data Skeptic

Fooling Computer Vision

Jan 22, 2020 · 25 min · 24.4 MB
0:00-25:26

Streams straight from the publisher. podnod never proxies or re-hosts episode audio.

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.