I agree with @terence.stone25 ...about the answer...and with @steve1 about the fact it's hard to wrap your head around because of other ways we talk about binary and the fact we start counting at zero.
So here's an esoteric response: zero is weird, man. Depending on the context it might describe nothingness, but it also happens to be the first non-negative integer which we use for counting or enumerating elements.
Here's a completely non-CS analogy: in Europe they often call the first floor of a building, floor 0. But that doesn't mean there is no floor there! It means they are enumerating floors starting at zero. If that's the only floor in the building we would say: this building has one floor - floor 0.
Back to CS: the first bit (upper left bit) of the image itself is at coordinate 0,0. But that doesn't mean there is no pixel there! There is 1 pixel. A 1x1 image has 1 pixel and that pixel is the zeroth pixel in the image.
The meta-data describes the size of the image not the set of values of possible coordinates.
@steve1 your last statement:
By this logic then
0000 0000 x 0000 0000 should be 1 x 1 which is tough to get my head around
Also by this logic we have no way to describe an image with zero pixels. Here's this for mind blowing: using this protocol it takes 16 bits to say the image has zero pixels.
It could be that someone decided, you know what, we don't need to have an image file format that lets you create a zero-pixel image (easier to just not have a file at all!), but they didn't. And this is actually common - the metadata of anything (even say an IP packet) should be able to say: this thing has nothing in it! And it takes some number of (metadata) bits to say that. Mind. blown.
For a hyperbolic example: I just saved a blank MSWord document on my computer and looked at the file size -- 21,649 bytes (or 173,192 bits). That's a lot of bits to say nothing is there.
Does that help?