Image and Video Recognition: A Bump in the Road
March 24, 2015
I read “Images That Fool Computer Vision Raise Security Concerns.” I found the write up a reminder that the marketing and venture capitalists’ hype are one thing. Real world software performance is another thing.
The article states:
Cornell researchers have found that computers, like humans, can be fooled by optical illusions, which raises security concerns and opens new avenues for research in computer vision.
The passage I highlighted in a mellow yellow says:
But computers don’t process images the way humans do, Yosinski [a Cornell wizard] said. “We realized that the neural nets did not encode knowledge necessary to produce an image of a fire truck, only the knowledge necessary to tell fire trucks apart from other classes,” he explained. Blobs of color and patterns of lines might be enough. For example, the computer might say “school bus” given just yellow and black stripes, or “computer keyboard” for a repeating array of roughly square shapes.
So what?
It turns out that this diagram looks exactly like a penguin.
The smart software sees the abstraction as what most grade school children know as a lovable penguin. I did not smell a penguin until after I left grade school. Someone should have warned me.
And the challenge? I have no comment about the expectations of a government professional who relies on image recognition as part of an on going investigation.
Stephen E Arnold, March 24, 2015
Comments
-
- Subscribe to Beyond Search
Feature archive
News archive
-
