[HN Gopher] Detecting edges of images at the speed of light
       ___________________________________________________________________
        
       Detecting edges of images at the speed of light
        
       Author : bookofjoe
       Score  : 82 points
       Date   : 2025-01-26 23:28 UTC (3 days ago)
        
 (HTM) web link (phys.org)
 (TXT) w3m dump (phys.org)
        
       | esperent wrote:
       | Edge detection is used a lot in 2d & 3d graphics, microscopy,
       | telescopy, photography, and a ton of other fields.
       | 
       | If this could be done almost instantaneously and without
       | additional energy, it would be very useful.
       | 
       | I can imagine having tiny analogue edge detection component in
       | cameras, microscopes, telescopes, and maybe even graphics cards.
       | 
       | This could be a part of the future of computing - tiny
       | specialized analogue components for specific tasks.
       | 
       | On the other hand, maybe digital compute, whether silicone or
       | futuristic alternatives, will always be cheap enough that the
       | economics will never work out - it'll always just be cheaper to
       | throw more general purpose compute at a problem.
        
         | 0u89e wrote:
         | I'm not really sure how any of this applies to software
         | development, as they detected edges in actual physical films,
         | which have different layers of chemicals. Still a very
         | interesting approach and that knowledge probably can be
         | transfered to other fields, as humans consist of layers of
         | chemicals, that are different from cars, houses and trees for
         | example, but could you and other people just actually read the
         | paper before writing anything here?
         | 
         | PS The title of article is on a clickbait level, as the meaning
         | for images nowadays does not correspond to what is used in
         | article.
        
         | kidel001 wrote:
         | To add onto this... some amount of what they are calling edge
         | detection here seems to overlap with what has already been
         | implemented in microscopy using phase contrast... which has
         | been around since 1932.
        
         | juancn wrote:
         | They made a 7 layer film that does edge detection photonically,
         | original paper:
         | https://pubs.acs.org/doi/10.1021/acsphotonics.4c01667
        
       | mcherm wrote:
       | Unless I am missing something, the article doesn't actually
       | explain HOW this was done, nor does it link to a research paper
       | or other source that explains it.
       | 
       | In my mind, this disqualifies this article from being something
       | "that good hackers would find interesting" and thus not
       | appropriate for this site.
       | 
       | (Please someone, tell me I missed the part where they actually
       | explained it.)
        
         | owobeid wrote:
         | The link to the paper is in the first paragraph.
         | 
         | https://pubs.acs.org/doi/10.1021/acsphotonics.4c01667
        
           | DarmokJalad1701 wrote:
           | "Tamm Plasmon Polaritons" sounds like something from Star
           | Trek technobabble.
        
             | nelblu wrote:
             | Darmok and Jalad on the ocean!
        
               | Galatians4_16 wrote:
               | Shaka, when the walls fell.
        
             | TeMPOraL wrote:
             | Technobabble in TNG-era shows was actually quite consistent
             | and even made sense if you squinted at it.
        
               | mrandish wrote:
               | Yes, there were people on the production staff who, in
               | addition to other duties, were responsible for
               | maintaining as much consistency as possible in both
               | technical descriptions and capabilities with prior
               | episodes and Star Trek canon. Mike Okuda was one such
               | person.
               | 
               | Obviously, the show was fiction and made for a broad
               | general audience, so their technical inputs were
               | sometimes 'streamlined' by the producers to minimize too
               | much technical explication dialogue. But they generally
               | did try to get enough on-screen to be both series-
               | consistent and technically plausible within the fictional
               | universe.
        
       | fzimmermann89 wrote:
       | If I am not mistaken, this is done by modulation in Fourier
       | space. We have already been using this in optical setups for ages
       | - at the speed of light.
       | 
       | The interesting part imo is the implementation of this idea in
       | their work and the efficiency and physical size.
        
       | TeMPOraL wrote:
       | Damn, there goes my idea. The other day I wondered how far you
       | could go by embedding classical and neural filters in plastic
       | sheets, to make a "magic lens" that does some img2img
       | transformation in real time. E.g. pixelizing should be relatively
       | simple, but I wonder if you could go as far as transforming a
       | small image classifier (say, "cat" vs. "hot dog") into a physical
       | object, to create a passive "lens" that turns incoming light into
       | words ("cat", "hot dog", "N/A")? Feels like it should be doable.
       | 
       | Anyway, thanks for the link; I'll study the paper and look up at
       | the references.
        
         | koromak wrote:
         | I suppose if your lens is actually a vertical stack of lenses,
         | each with hundreds of inputs and outputs, then why wouldn't
         | this work? Although I cannot fathom fabricating it. Maybe start
         | by finding the absolute simplest/smallest image classifier you
         | can
        
           | koromak wrote:
           | EDIT: Of course people are already doing this. D2NN seems to
           | be the keyword, Diffractive Deep Neural Networks
        
             | TeMPOraL wrote:
             | Yes, by "lens" I meant a possibly large stack of optical
             | elements, each designed to perform a single computational
             | step. Kind of like layers of an NN, but etched on sheets of
             | plastic.
             | 
             | Thanks for finding the proper term for this!
        
         | Xmd5a wrote:
         | I've had the same idea for a while, but the classifier I had in
         | mind was for identifying wood ("yep, it's wood"). I wonder if
         | this kind of tech could be used for low-light enhancement
         | (night vision). This would require "injecting" light into the
         | "lens" I guess.
         | 
         | MNIST classifier using a DNN:
         | https://www.nature.com/articles/s44172-024-00211-6
        
         | Someone wrote:
         | > pixelizing should be relatively simple
         | 
         | https://www.yankodesign.com/2024/07/04/this-crystal-
         | fragment...:
         | 
         |  _"The Pixel Mirror measures 16mm x 16mm x 10mm, which means
         | it's small enough to be worn as a pendant in a necklace.
         | Monoli's series of wearable and handheld prisms are all
         | handmade, and because of the nature of polishing natural
         | stones, they are not perfect square "pixels". They are handmade
         | to suit the condition of the available stone."_
         | 
         | Discussed on HN in
         | 
         | https://news.ycombinator.com/item?id=40895907 and
         | https://news.ycombinator.com/item?id=40890633
        
       | infogulch wrote:
       | Huygens Optics has a great video "Fourier Optics used for Optical
       | Pattern Recognition". He starts with explaining the Fourier
       | Transform wrt audio, then demonstrates a physical filter that
       | screens the resulting convolution pattern to identify matches.
       | Very cool: https://www.youtube.com/watch?v=Y9FZ4igNxNA
        
       | uhahohheck wrote:
       | OT, but: Oh...wait...just another thursday, ich came here to HN
       | just for the new games, and no... not one single ...no new game
       | at all on the frontpage of HN... so what the heck, busy everyone
       | ?
       | 
       | [modus:with.a.sense.for.the.weekend]
       | 
       | regards,...
        
       | kidel001 wrote:
       | Technically... they are outlining edges at the speed of light.
       | Detection is a separate process entirely.
        
       ___________________________________________________________________
       (page generated 2025-01-30 23:01 UTC)