How does an eye reconstruct an image?

[Salmone]

I would like to understand the relationship between the image formed on the retina and the image reconstructed by the human eye. Let's consider this images:

The object B forms the real image OB' while the object A forms the real image OA' on the retina. Second image:

here, the object forms a virtual image $h_i$. My question is: what is the correlation between the image defined in geometrical optics as the the point where the emerging rays converge and the image recreated by our brain? For example, in the second image, there is an image formed on the retina and a virtual image $h_i$, where do we see the object? I think it is $h_i$, if so, where do our brain reconstruct the images of the first picture? The point where cones and rodes are hitted by light on retina, what kind of informations gives us about where to reconstruct the image?

Like: rays of light start from the arrowhead A and travel to the eye, here there is a lens which focuses all the rays into a certain point on the retina depending on where the rays started from; what happens then? I don't want an answer related to medicine of course, I just can't understand correlation between image on the retina and image our brain let us see.

 

[andrewkirk]

where do our brain reconstruct the images of the first picture

Unless you are asking what part of the brain processes the image (and I don't think you are asking that) this question has no answer. It is like asking "where is my sadness?"

The brain uses the information it gets from the pixel pattern on the retina to construct an experience for you, and that experience is what we call "seeing". It has a spatial or geometric nature to it and makes us feel that the object we see is "over there" (and countless philosophical tomes have discussed and debated what we mean by "over there"). But the experience is not an object in space, any more than other experiences such as sadness, fear or curiosity are.

Even pain in a specific part of the body, that we think of as very spatially specific - eg to the left knee, is actually an experience created by our brain, based on signals it receives through nerves from that part of the body. That's why local anesthetic techniques such as "ring blocks" work. If the signal from your finger where the doctor is digging out a deeply embedded splinter, can't get to the brain, you feel no pain.

 

[DaveC426913]

countless philosophical tomes have discussed and debated what we mean by "over there"

It's everything that's there at the end of my eye feelers.

 

[tech99]

The image is just a flat pattern lacking distance information. The eye will judge the distance to an object in two ways. By adjusting the lens power using the ciliary muscles until it is sharpest, the brain knows the approx distance. In addition, by using binocular vision together with geometry, distance can be established using both eyes together. In some respects, the process of seeing is more like image processing than a traditional camera. The brain is looking at past experience and seeking clues to give a best estimate of the image. That is why optical illusions can fool us.

 

[Drakkith]

here, the object forms a virtual image hi. My question is: what is the correlation between the image defined in geometrical optics as the the point where the emerging rays converge and the image recreated by our brain?

Each spot on your retina is mapped to a single spot in your perceived visual field. So light that falls on point A' always looks like it is a certain spot in the top part of your visual field (the image is reversed and upside down, but your brain sort of flips it around during processing). Light that falls on B' will always appear between the center of your vision and point A'. Where this light comes from doesn't matter. If it falls on a certain spot on the retina, it will show up at a certain spot in your visual field.

 

[DaveC426913]

Where this light comes from doesn't matter. If it falls on a certain spot on the retina, it will show up at a certain spot in your visual field.

And you can hack it too.
Close your eyes and press on your eyeball in the outer corner of your eye.
You will see a bright patch of light in the inner corner of your vision.

 

[jbriggs444]

The image is just a flat pattern lacking distance information. The eye will judge the distance to an object in two ways. By adjusting the lens power using the ciliary muscles until it is sharpest, the brain knows the approx distance. In addition, by using binocular vision together with geometry, distance can be established using both eyes together. In some respects, the process of seeing is more like image processing than a traditional camera. The brain is looking at past experience and seeking clues to give a best estimate of the image. That is why optical illusions can fool us.

In my experience, binocular vision is the a more important depth cue than per-eye focus. The eyes will make an initial focus attempt at the depth suggested by binocular vision and will then "hunt" to find the depth which maximizes image sharpness. It works just like a camera. By interfering with the binocular cue, I've been able to watch as my eyes auto-focus.

However, binocular vision is not the only distance cue and is not always the most important. Image size, vertical position and occlusion also enter in.

Edge detection is important. This is how we infer shape in stick images. The visual cortex finds edges. It looks for places where edges meet in a corner and makes a guess about whether this is a convex corner (corner facing us) or a concave corner (facing away). It then does a local comparison with surrounding corners on the associated edges to make a global assessment on the shape of the perceived object.

One of my professors in college in the 70's was seriously into image analysis on this basis. He gave us a once over one afternoon.

It is my impression (perhaps false) that there is a two dimensional spatial distribution in the visual cortex corresponding to the spatial distribution of rods and cones on the retina. This is similar to how touch on the skin is mapped physically to a two-dimensional layout in the brain.

 

[berkeman]

Close your eyes and press on your eyeball in the outer corner of your eye.

I think it's generally a bad idea to press on your eyeballs like that. I'll have to look for a valid reference, but my ophthalmologist has mentioned it to me in the past.

My question is: what is the correlation between the image defined in geometrical optics as the the point where the emerging rays converge and the image recreated by our brain?

I think an important point about the retina that has not been mentioned yet is the varying level of optical resolution with radial distance away from the fovea. The resolution of rods and cones is highest near the fovea, and drops off significantly with distance outward. This means that our brains are filling in resolution for those areas of lower sensing resolution. Amazing stuff.

https://en.wikipedia.org/wiki/Fovea_centralis

 

[DaveC426913]

I think it's generally a bad idea to press on your eyeballs like that. I'll have to look for a valid reference, but my ophthalmologist has mentioned it to me in the past.

Then you really don't want to read about Newton's experiments using a darning needle!

 

[tech99]

The evolution of nerve endings to respond to light has parallels with the original use of a shift register as an optical scanner (hence the mobile phone camera). And the very computational way the brain sees images has parallels to the speech processing used by mobile phones, where we are only presented with an illusion of the voice.