You can derive it from the fact that there's no feasible way to get such uniform (and cute) image data for every food item.

Though I want to add that this is a good application of AI image gen since the images are useful for quick visual confirmation that the item is in the same ballpark of the thing that you think it is.