Focus - Plane or Point

Tiziano

I'm writing a paper and I need to know the differences between how a camera focuses and how our eyes focus.

My guess is that our eyes focus on a point (well, two slightly divergent points). Therefore, things get progressively more out of focus as they get farther away from the focal point, spherically.

And a camera focuses on a plane. Everything within that plane is in equal focus. Everything farther away from the plane gets progressively more out of focus depending on how far away from the plane it is.

Am I correct?

pathfinder

Sounds pretty reasonable.

The main difference between a camera and an eye, is that the acuity - sharpness of the image - in an eye is only the central 2-4 degrees - everything beyond the central 2-4 degrees of central field is quite blurred. Whereas, as you said, with a camera, everything in the image circle is pretty sharp - yes, the center is sharper than the extreme corners, but there in no where near as much loss in sharpness as in an eye.

Altho the eye is frequently described as a camera, I think a much better desription is that of a real time scanner, as the "image" we believe we see is NOT what the eye is actually seeing real time.

What we "see" is an assembling of numerous visual scans that is assembled by the visual cortex and the conscious mind. For example, almost no one is aware of a blind spot in our vision, but there is a permanent blind spot about 15 degrees temporal to central vision in all of us. That is pasted over in our mind's eye, so we are not aware of it.

ivar

pathfinder wrote:

Altho the eye is frequently described as a camera, I think a much better desription is that of a real time scanner, as the "image" we believe we see is NOT what the eye is actually seeing real time.

What we "see" is an assembling of numerous visual scans that is assembled by the visual cortex and the conscious mind. For example, almost no one is aware of a blind spot in our vision, but there is a permanent blind spot about 15 degrees temporal to central vision in all of us. That is pasted over in our mind's eye, so we are not aware of it.

the sensor/fllm is the same all over, while with the eye you have to deal with the fovea and the blind spot. There are a bunch of blind spot tests around on the internet, this is one of them. (See paragraph 'blind spot').

i think you are on the right though as far as the focus thing goes

luke_church

pathfinder wrote:

The main difference between a camera and an eye, is that the acuity - sharpness of the image - in an eye is only the central 2-4 degrees - everything beyond the central 2-4 degrees of central field is quite blurred.

Absolutely. The comparison between the eye and a camera is common, though often very confused one.

There are several points to note here:

-> The central point in the eye (the fovea), has a very high spatial acuitity .

-> The rest of the eye trades spatial acuitiy for sensitivity, by accumulating multiple 'sensor cells' into one. In this way it achieves a sensitivity that camera manufactures can only dream of; in very dark environments, close to single-photon levels. I think this trade-off becomes increasing strong as you progress away from the centre of the retina.

-> The retina is a small, and non-flat sensor, so when addressing whether it's 'sharp' it's kind of important to be clear about what you're saying. Whether something is optically sharp is different from whether it's perceptually sharp. I would be hesitant to make claims about the operations of the human lens not resolving to a plane.

Sure it won't literally resolve to a plane, it'll be a much more complex shape like a real camera lens, and will have fall off effects etc. but I'm not so sure that the loss of sharp-ness isn't due to the pre-attentive visual processing, such as grouping of rods.

>Altho the eye is frequently described as a camera, I think a much better desription is that of a real time scanner, as the "image" we believe we see is NOT what the eye is actually seeing real time.

Again, absolutely. It's a well quoted statistic that there are an order of magnitude more neural connections from the brain to the eye than the other way round. What we 'see' is the result of a hugely complex set of processing operations, for example it is believed that your eyes implicitly have the ability to do sophesticated edge detection, pattern recognition, missing information inference, and even sub-coding into Gabor Wavelets (think, really funky mathematical encoding of the image)...

There have been experiments on video compression done where you detect what the person is 'fixating on' and gaussian blur the rest away, people didn't recognise the difference.

It seems that world knowledge is possibly more important that the actual light sources that are reaching the eye.

It's a really fun topic....

If you're interested or looking for references for your essay:

-> The standard image processing text by Gonzalez and Woods talks quite a lot about it.

-> http://www.cl.cam.ac.uk/teaching/2004/Graphics/camonly/cgip1.pdf

(Pretty notes, some of which are taken from Gonzalez, p7 has a few cute diagrams)

-> http://www.cl.cam.ac.uk/teaching/2006/CompVision/CompVisNotes.pdf

More serious notes about the importance of image processing in human vision. Just skip over the maths stuff if you're not that way inclined, it highlights a lot of the problems that most of us just take for granted :-)

Enjoy. Hope this helps.

PS. You'll find a lot of this stuff is often considered to be 'cognitive psychology'. CogPsy people have 'done the eye to death' in research levels, often in a way that was unfortuante for kittens :-(, but on the plus side, Google for that will probably find you as much as you ever wanted to know... The keyword 'Gestalt' might also help you cut out some of the dross...