4 ms·

I was just wondering a couple days ago why the image from my phone is so grainy, while my eyes+brain can see everything clear in the dark (it wasn't completely

I was just wondering a couple days ago why the image from my phone is so grainy, while my eyes+brain can see everything clear in the dark (it wasn't completely dark, of course). This seems to replicate the post-processing we do in our brain (which is also a giant neural network). I wonder if the process is similar?

jacobush 7y ago
Also, our eyes are better. So far.
- xwdv 7y ago
  I wonder what things would be like when phone cameras are as good as human eyes, the climax of consumer photography?
koverda 7y ago
That's not really a good analogy. You have a totally different sensor chemistry in your eyes, as well as different processing.
- ganstyles 7y ago
  And while brains are the original neural networks, they don't resemble what's going on with ML DNNs at all.
  - arketyp 7y ago
    Low-level visual processing in the cortex share striking similarities with CNNs, actually. Gabor filters etc.
arpa 7y ago
Very small numbers of photons (1) are required to trigger rhodopsin cycle. So primary receptor itself is very very VERY sensitive.
- tofof 7y ago
  To be clear, the parent did not fail to include a citation. The parenthetical note is that rod cells are so sensitive that they react to being struck by a single photon.
- folli 7y ago
  That's actually a very cool fact: https://www.nature.com/news/people-can-sense-single-photons-1.20282 https://www.nature.com/news/people-can-sense-single-photons-...
empath75 7y ago
Your brain doesn’t make a 2 dimensional image based entirely on photons entering your eye. You generate a complex physical model of your surroundings based only partially on visual input and rely substantially on memory.
- zaroth 7y ago
  Kind of like this well trained CNN is no longer relying entirely on the raw pixel values, but is statistically inferring a brighter image from the baseline.
  - jdmichal 7y ago
    There's a difference between applying known priors, and making things up based on statistics. Conflating the two isn't helping anyone.
    - zaroth 7y ago
      Not to harp on this, but the point is that, as I understand it, both “systems” are using exogenous information to extrapolate more data than is actually present in the source image. That’s not to say that the same “thing” is happening at the granular level at all. But this is distinctly different from standard filtering functions, which can only work with entropy already present in the source image. So there’s a neat distinction. The output from the CNN is essentially an “artist interpretation” of the source image. As such there could be “clarifying details” in the output that were in fact totally invented and not actually present in the source.
- jdmichal 7y ago
  Also other senses, including proprioception. In a completely dark environment, you could swear that you see your hand waving in front of your face. That's because your brain actually does know it's there, and it's trying to create a unified model.
Someone 7y ago
The dynamic range of the human eye is better than that of the lens in your phone. https://en.wikipedia.org/wiki/Human_eye#Dynamic_range https://en.wikipedia.org/wiki/Human_eye#Dynamic_range: ”The human eye can detect a luminance range of 10¹⁴, or one hundred trillion (100,000,000,000,000) (about 46.5 f-stops), from 10−6 cd/m2, or one millionth (0.000001) of a candela per square meter to 10⁸ cd/m2 or one hundred million (100,000,000) candelas per square meter. This range does not include looking at the midday sun (10⁹ cd/m2)[21] or lightning discharge.”

jacobush 7y ago

Also, our eyes are better. So far.

xwdv 7y ago
I wonder what things would be like when phone cameras are as good as human eyes, the climax of consumer photography?

koverda 7y ago

That's not really a good analogy. You have a totally different sensor chemistry in your eyes, as well as different processing.

ganstyles 7y ago
And while brains are the original neural networks, they don't resemble what's going on with ML DNNs at all.
- arketyp 7y ago
  Low-level visual processing in the cortex share striking similarities with CNNs, actually. Gabor filters etc.

arpa 7y ago

Very small numbers of photons (1) are required to trigger rhodopsin cycle. So primary receptor itself is very very VERY sensitive.

tofof 7y ago
To be clear, the parent did not fail to include a citation. The parenthetical note is that rod cells are so sensitive that they react to being struck by a single photon.
folli 7y ago
That's actually a very cool fact: https://www.nature.com/news/people-can-sense-single-photons-1.20282 https://www.nature.com/news/people-can-sense-single-photons-...

empath75 7y ago

Your brain doesn’t make a 2 dimensional image based entirely on photons entering your eye. You generate a complex physical model of your surroundings based only partially on visual input and rely substantially on memory.

zaroth 7y ago
Kind of like this well trained CNN is no longer relying entirely on the raw pixel values, but is statistically inferring a brighter image from the baseline.
- jdmichal 7y ago
  There's a difference between applying known priors, and making things up based on statistics. Conflating the two isn't helping anyone.
  - zaroth 7y ago
    Not to harp on this, but the point is that, as I understand it, both “systems” are using exogenous information to extrapolate more data than is actually present in the source image. That’s not to say that the same “thing” is happening at the granular level at all. But this is distinctly different from standard filtering functions, which can only work with entropy already present in the source image. So there’s a neat distinction. The output from the CNN is essentially an “artist interpretation” of the source image. As such there could be “clarifying details” in the output that were in fact totally invented and not actually present in the source.
jdmichal 7y ago
Also other senses, including proprioception. In a completely dark environment, you could swear that you see your hand waving in front of your face. That's because your brain actually does know it's there, and it's trying to create a unified model.

Someone 7y ago

The dynamic range of the human eye is better than that of the lens in your phone. https://en.wikipedia.org/wiki/Human_eye#Dynamic_range https://en.wikipedia.org/wiki/Human_eye#Dynamic_range: ”The human eye can detect a luminance range of 10¹⁴, or one hundred trillion (100,000,000,000,000) (about 46.5 f-stops), from 10−6 cd/m2, or one millionth (0.000001) of a candela per square meter to 10⁸ cd/m2 or one hundred million (100,000,000) candelas per square meter. This range does not include looking at the midday sun (10⁹ cd/m2)[21] or lightning discharge.”