Edge detection

Discussion in 'Intelligence & Machines' started by domesticated om, Jul 4, 2010.

Thread Status:
Not open for further replies.
  1. Billy T Use Sugar Cane Alcohol car Fuel Valued Senior Member

    Messages:
    23,198
    That is your false assertion, given without proof, or even a reply to post 19, which explained to you that EVERY NERVE DISCHARGE PRODUCES MORE ENTROPY.

    I.e. the entropy reducing work of the nerve cell's "Sodium Pump" is destroyed by nerve discharges just as the entropy reducing work of men neatly stacking up a tower of brick is destroyed when they are pushed over.

    Did you even read post 19? - It clearly explains that all of brain information processes, such as nerves firing, are entropy producing. They destroy the reduction of entropy made by the sodium pumps of the brain cells.

    BTW I discussed, in post 7, the fact that a great deal of processing does occurs in the retina, noting that it had to as the information transfer capacity of the optic nerve is about 50 times less than the rate retinal cells collect information. I agreed that much of this is closely related to edge detection but I tend to be more general and call it "data compression" as there is more to it than just edge detection AND the edge detection of the retina does not have any orientation data with it. Only at in that region nearby cells have different levels of illumination. "Edge detectors" in V1 (and the fact they are orientation sensitive) is why Hubble and Wessel got the Nobel prize.

    Most workers in neurology hold my POV that edge detection is the process H&W described in V1 and the retinal processing is data compression, even though much of the data that makes it thru the optic nerve is where their are contrast boundaries (edges). The difference it that the retina data transmitted does not "tag" the edges with information about the orientation of those contrast areas as V1 does. I further discuss the important fact that these orientation sensitive edge detectors in V1 mutually interact in V1 with nearby other edge detectors. I.e. like orientation edge detectors mutually stimulate each other whereas orthogonal ones mutually suppress each other.

    Read my discussion of this in post 9, whcih show how this mutual interaction between V1 edge detectors allows the CONTINUOUS visual field of stimulation to be parsed into objects. (There are no "objects" separated one from another in the retina. That is built on V1's edge detectors mutual interactions.)
     
    Last edited by a moderator: Nov 22, 2010
  2. Google AdSense Guest Advertisement



    to hide all adverts.
  3. Blindman Valued Senior Member

    Messages:
    1,425
    I did read post 19 as it evolved.. The brain is not a closed system. If it where you would be correct. But it is not and though the entropy of the universe increases with brain function the entropy of the brain is kept low by increasing the entropy outside the brain. Energy can be used to decrease entropy, increasing complexity, but is sacrificial via increasing entropy else where..
     
    Last edited: Nov 22, 2010
  4. Google AdSense Guest Advertisement



    to hide all adverts.
  5. Billy T Use Sugar Cane Alcohol car Fuel Valued Senior Member

    Messages:
    23,198
    Yes it is. In the brain it is the cell's "sodium pumps" that reduce entropy. They use the huge energy the brain requires. (I.e. they fight diffusion to make the non-equlibrium condition with the interior of the cells at -70mV and having much lower Na+ ion concentrations than outside.)

    It is the nerve discharges that DESTROY THIS STATE = INCREASE THE ENTROPY = RESTORE THE EQUILIBRIUM CONDITION, AS DIFFUSSION WOULD DO ALSO.

    Fact that heat is transfered to the skin, etc. by blood flow keeps the brain from over heating, but that is not "reducing entropy." - That is reducing temperature. Only the sodium pumps reduce entropy in the brain. ALL brain activity related to thought, etc, INCREASES entropy!!!! "Thinking" etc. is a thermodynamically IRREVERSIBLE process so is entropy producing.

    Thus, at least we agree that the more you think, the more you increase the entropy of the universe. I.e, the brain is the body's greatest entropy producer, pound for pound. - I don't see how you can (still without any proof) claim it is reducing entropy, is the lowest entropy organ, etc. Quite the opposite is true.

    later by edit: I agree with what you state in post 22. (Except I would be even more correct, if the brain were a closed system -i.e. correct in either case - until the brain overheated and died. Even then, as it decayed entropy would be increased as that too is an irreversible process.)
     
    Last edited by a moderator: Nov 22, 2010
  6. Google AdSense Guest Advertisement



    to hide all adverts.
  7. Blindman Valued Senior Member

    Messages:
    1,425
    Please just post anew rather then add paragraphs to old posts. I makes it hard to discuss a topic if your posts continue to change each time I read them. Use edit to correct spelling and grammar not to add multiple paragraphs.

    Anyways. I totally agree that edge detection is not the only function of the neural net in the retina.

    Orientation is important but you misunderstand that it is relative. Retinal networks can be shown to be sensitive to specific orientations.

    It is important to understand that the system has back propagation. It modifies its self via feedback. Thus if I rotate the image, via lenses to flip the field of view, optic nerves responding to a particular orientation the same way once it has learnt the anomaly in input. This has in the past been described as orientational indifference, an incorrect conclusion.
     
  8. Blindman Valued Senior Member

    Messages:
    1,425
    OMG you just removed 2 paragraphs from post 23 and added two new ones.. Please please make it linear..
     
  9. Billy T Use Sugar Cane Alcohol car Fuel Valued Senior Member

    Messages:
    23,198
    I have mild dyslexica. I read what I want to be there, not what I typed. Strangely after it is posted in a different format, I can often see my typos and ommission. When correcting them I do often add more, but usually correct without removals.

    I do not know of any support for back transmission in the optic nerve. Are you stating there is?** If so why? I know a lot of back transmission does exist in the Brain and use some of these little know facts to support my POV about how perception works.

    For example you will rarely find any discussion of the fact that slightly more nerve fibers come to V1 (start of the visual cortex) from the parietal brain than from the eyes via the LGN, because in the accepted theory of perception, they have no reason to exist.

    They are essential and predicted by my theory of how perception works. (I.e. it is a creation of parietal brain, not the "emergent" end result of many stages of forward flowing neural computational transforms. To keep the parietal Real Time Simulation* faithful to external reality, it must send its constructed "characteristics" back to V1 for comparison with the input coming from the eyes.)

    *To quickly understand some serious flaws of the accepted theory of perception, I hope you will first read: my short post at: http://www.sciforums.com/showpost.php?p=2502342&postcount=12

    Then to understand my alternative POV about perception read my longer “RTS essay” at: http://www.sciforums.com/showpost.php?p=905778&postcount=66

    **Perhaps you are only stating that the visual system does not care about the orientation of the retinal images (but each eye must be the same for reason I won't go into to). That is true, whatever you have been using is OK. With inverting lenses you can flip the retinal image to be upright instead of the normal inverted. One guy got so good adapted so well, to invering lenses that in a few weeks he rode his bike thru traffic with them on!
     
    Last edited by a moderator: Nov 22, 2010
  10. Blindman Valued Senior Member

    Messages:
    1,425
    Do you read what I type??? I AM SAYING THAT THE BODY KEEPS THE BRAINS ENTROPY LOW BY INCREASING ENTROPY ELSEWHERE.....

    How could we possibly function if we burnt our brains at our first thoughts. Our brains can function for 100+ years. That is approx 46Wh a day, 16.4KWh a year and well over 1MWh in a lifetime. Please do explain how a 1.5kg brain can burn that much energy without finding a way to move the entropic imbalance away. And some lucky people are just as functional at 100 as at 20...
     
  11. Billy T Use Sugar Cane Alcohol car Fuel Valued Senior Member

    Messages:
    23,198
    Yes you said that in post 22 & I agreed with you as last part of post 23, perhaps added after your had already read 23?

    I agreed with this earlier too. cooling of the brain is essential. - that was the main point of largest paragraph in post 23.

    This, your first post, is what I reject:
    Most of the energy used by the heart is for a reversible pumping process. ALL of the nerve firing in the brain are irreversible. Reversible processes do not produce entropy.
     
    Last edited by a moderator: Nov 22, 2010
  12. Blindman Valued Senior Member

    Messages:
    1,425
    I am not fully aware of the amount of back propagation via the optic bundles. I know that there is very strong back propagation in the retinal layers. There are several nodes into the brain along the optic bundles each with many layers and alot of back propagation at each node. Edge detection is very early in the data propagation from the eye
     
  13. Blindman Valued Senior Member

    Messages:
    1,425
    So we can finally agree that the complexity of the brain is higher then other organs because of its ability to do work. It is a very efficient system that maintains low entropy and thus high complexity via many systems that move simplicity away from the body.
     
  14. Blindman Valued Senior Member

    Messages:
    1,425
    WTF I cant communicate with you and your constant edits. Post 28 just changed with a complete change of the final paragraph and quote...
     
  15. Billy T Use Sugar Cane Alcohol car Fuel Valued Senior Member

    Messages:
    23,198
    To Blindman:

    Part of what has misslead you into making false statements in your first post (16) is your idea, there in stated: “Information processing requires lower entropy.”

    Do you remember the old Cray computers? At one time the world’s fastest and most powerful. They were huge energy hogs and great entropy producers. They achieved their status because they could be cooled. – They were a sealed box with a fluid (Freon?) circulating around all the components. They were fast because the parts were close together. Many modern computers have high density and get hot – Irreversibly produce a lot of heat and entropy.

    SUMMARY: Information processing does NOT inherently require low entropy processes, but it will require a lot of energy, if it is not done reversibly.
    No computer is reversible - To be that it would need to never erase any part of the prior process. Thus, ALLcomputers produce a lot of entropy, the brain included. In contrast, the heart produces little entropy as it is mainly a pump - REVERSIBLE use of energy.
     
    Last edited by a moderator: Nov 22, 2010
  16. Blindman Valued Senior Member

    Messages:
    1,425

    You have to understand that the maintenance of complexity in a system is based on the ability of the system to move entropy away.

    I can start a camp fire and it will quickly burn out. I can maintain that fire by supplying more fuel but that process requires work. From the process that grows the wood to the energy required to move the wood onto the fire. But as long as I supply the fuel the fire has a low entropy and the ability to do work. It maintains its complexity.

    A system must maintain low entropy to do work. Entropy also is a measure of complexity.

    So as you say the heart uses much less energy then the brain, it intrinsically can function at higher entropy, hence it is less complex
     
  17. Blindman Valued Senior Member

    Messages:
    1,425
    It would be wonderful if you where right. Perpetual motion. All work increases entropy but this can be moved into other parts of the system (universe).

    You have to love radioactive isotopes as they have exchanged vast amounts of entropy into the universe so that they can do lots of work billions of years latter.
     
  18. Billy T Use Sugar Cane Alcohol car Fuel Valued Senior Member

    Messages:
    23,198
    I obviously do. - I just gave you the example of the old Cray computer and noted that in it computation was a high entropy production process (A counter example to your still undefended false post 16 claims.) because it could move the great entropy and heat produced away with fluid cooling. It is sort of insulting for you to imply that I don't understand this. I stated that if the brain were an isolated system it would over heat and die.

    At least here you seem to be reversing your original claim that complex processes like brain computing must be low entry processes. Anyway you should as I have shown that the firing of every nerve is producing entropy in irreversible process and that NO computer is reversible and noted that in contrast the heart, as a pump, is reversibly using energy. (Reversible proceses produce zero entropy.)

    Have you forgotten what you claimed to the contrary in your original post 16?
    You have morphed that into notes that heat and entropy must be carried away - and I have repeatedly agreed to that. Again my agrument with you is that your original post 16 has several false statements.

    (1) Complexity is not measured by the pound.
    (2) The brain is not lower entropy organ than the heart.
    (3) Computing does not require a low entropy device.
     
    Last edited by a moderator: Nov 22, 2010
  19. Billy T Use Sugar Cane Alcohol car Fuel Valued Senior Member

    Messages:
    23,198
    I never claimed that any process was fully reversible, but one can recover a large part of the energy driving well designed pumps. For example many cars did this in the "fluid drive systems."

    This is in stark contrast to the brain's computation processes - they (the firing of nerves) are 100% irreversible - heat and entropy producers. This is as I have stated why the brain is a much greater producer of entropy than the heart.

    I will agree that the entire circulation system is dissipative but you spoke of the heart, not the capillary bed.
     
    Last edited by a moderator: Nov 22, 2010
  20. Blindman Valued Senior Member

    Messages:
    1,425

    I think you should re think your post as this is what I said in regard to your 3 points

    (1) "Complexity is not measured by the pound." and my post.
    (2)"The brain is not lower entropy organ than the heart." and my post.
    (3)"Computing does not require a low entropy device." and my post.
    Dude I am BE, BSc, PhD. I have no problem arguing in a debate but you have to be careful with your interpretation of my posts.
     
  21. Blindman Valued Senior Member

    Messages:
    1,425
    What, a nerve can only fire once??????

    Yes the brain does increase entropy more then the heart, but to do this it requires more complexity...

    We are running in circles and its 4:30am in Western Australia.. Tomorrow I would love to continue but now good night... I love the debate

    Mark
     
  22. Billy T Use Sugar Cane Alcohol car Fuel Valued Senior Member

    Messages:
    23,198
    No. Some with strong Na+ pumps can restore the -70mV internal "resting potential" in less than 0.001 sec. - i.e. fire at 1000 Hz, but most fire at 500 Hz or less. Most believe it is the change in firing rate that conveys information (and I agree) but I also suspect that the interval between pulses may too. Very closely related, for individual nerves, but in a network they could tell different things.

    I can understand why you asked. I did say the firing was 100% irreversible. What I meant by that was the the Na+ ion's reduction in entropy was 100% destroyed by a firing. I.e. the Na+ pump must make the entire entropy reduction again. (Or almost that if nerve can fire again before the internal potential is fully restored to -70mV. - Perhaps the depolarization wave that travels down the axon can be triggered when the internal voltage is only -60 mV?)

    My analogy of team of men building a tall tower of bricks from a random pile of bricks (reducing entropy) and then having some one knock it over (rapidly increasing entropy) is really quite good aid to understanding where / when lots of energy is required and when entropy is rapidly increased with the Na+ pump being the team of men and the nerve firing being the guy knocking the tower over. Main difference is the men need a year to make 500 tall towers, but the nerve can make -70Mv and fire 500 or so times each second!

    True in this case but you seem to be falsely implying that complexity is needed to make a lot of entropy. Grinding canon bores is not very complex but makes a lot of entropy as the well organized casting is reduced to a pile of tiny chips. Entropy production is very weakly if at all related to complexity or the weight of the device. It is directly related to the irreversibility of the process. This is why I had problems with your post 16 statements which mentioned complexity and weight, but not irreversibility.

    It is 7:20PM here in Brazil so good night, sleep well, We can continue if you like when I get up tomorrow.
     
    Last edited by a moderator: Nov 22, 2010
Thread Status:
Not open for further replies.

Share This Page