How Does a Computer Recognise a Cat?

A computer doesn't "see" a cat the way you do. It sees a huge grid of numbers — one number for the brightness of every tiny dot (pixel) in the photo.

To learn what a cat looks like, we show the computer thousands of photos already labelled "cat" or "not cat." Slowly, it starts to notice patterns: pointy ears here, whiskers there, a certain shape of eyes.

It doesn't memorise the photos — it learns the pattern. That's the difference between a computer that just stores pictures and one that can recognise a cat it has never seen before.

This process is called training, and the "testing" photos (ones the computer never saw during training) are how we check whether it actually learned the pattern, or just got lucky.

Try it yourself in the AI Hub's Train-the-Classifier game — you'll train a mini model in your browser and watch its accuracy improve in real time.