Computer Vision: a (very) brief intro
Image from a talk by Joseph Redmon presented at TED2017.

Computer Vision: a (very) brief intro

Helping machines 'see' using computer vision can be very beneficial to us. We can use computer vision to help blind people 'see', to diagnose cancer more effectively, and to teach cars to drive themselves.

It'll be exciting to see what other applications we come up with as the technology improves!

Computer vision works using machine learning algorithms and lots of labeled 'training' images. The algorithm looks through all the pixels in the training images and picks up patterns of what's what. Eventually, it gets good enough that you can show it an image it's never seen before and it'll know what's in it. This is kind of like how we learn to tell the difference between things.

When you were younger, you may not have been able to tell different dogs breeds apart. After seeing lots of examples though, you probably learned to tell the difference between a Labrador and a German Shepherd. That's basically how the computer learns, and the same method can be used to teach computers to 'see' almost anything.


要查看或添加评论,请登录

Mark Meleka的更多文章

社区洞察

其他会员也浏览了