Post

Humans are really good at recognizing an image even when it's been cropped, rotated, deep-fried, de-saturated, etc. Computers, in general, aren't as good: different bits=different image. Our latest dripline post discusses some techniques that allow computers to recognize similarity at scale