Lesson 1 of 4

How image AI creates a picture

15 pts

Explanation

Many modern image generators begin with visual noise: random-looking pixels. During training they learn how to remove a little noise while moving the image toward the description in the prompt. When you ask for an image, the model repeats this denoising process many times. It does not search a folder for one finished picture; it builds a new image step by step from learned visual patterns.

Visual explanation

Read this from left to right: what goes in, what happens, and what comes out.

Random noise
Denoise with prompt
New image

Each step removes a little random noise while following the prompt’s visual direction.

In plain language

Image AI starts with random visual noise and improves it in many small steps until it looks more like the requested image.

Remember

It creates a new image through many small changes.

What to do

Does an image generator usually begin with a finished photo, or with noise that it gradually turns into an image?

Validation questionWhat does a diffusion-style image model repeatedly remove to form an image?