TE
TechEcho
Home24h TopNewestBestAskShowJobs
GitHubTwitter
Home

TechEcho

A tech news platform built with Next.js, providing global tech news and discussions.

GitHubTwitter

Home

HomeNewestBestAskShowJobs

Resources

HackerNews APIOriginal HackerNewsNext.js

© 2025 TechEcho. All rights reserved.

DifuzCam: Replacing Camera Lens with a Mask and a Diffusion Model

116 pointsby dataminer9 months ago

14 comments

dekhn9 months ago
For those interested in various approaches to lens-free imaging, Laura Waller at Berkeley has been pursuing this area for some time.<p><a href="https:&#x2F;&#x2F;waller-lab.github.io&#x2F;DiffuserCam&#x2F;" rel="nofollow">https:&#x2F;&#x2F;waller-lab.github.io&#x2F;DiffuserCam&#x2F;</a> <a href="https:&#x2F;&#x2F;waller-lab.github.io&#x2F;DiffuserCam&#x2F;tutorial.html" rel="nofollow">https:&#x2F;&#x2F;waller-lab.github.io&#x2F;DiffuserCam&#x2F;tutorial.html</a> includes instructions and code to build your own <a href="https:&#x2F;&#x2F;ieeexplore.ieee.org&#x2F;abstract&#x2F;document&#x2F;8747341" rel="nofollow">https:&#x2F;&#x2F;ieeexplore.ieee.org&#x2F;abstract&#x2F;document&#x2F;8747341</a> <a href="https:&#x2F;&#x2F;ieeexplore.ieee.org&#x2F;document&#x2F;7492880" rel="nofollow">https:&#x2F;&#x2F;ieeexplore.ieee.org&#x2F;document&#x2F;7492880</a>
评论 #41280881 未加载
karmakaze9 months ago
This is not a &#x27;camera&#x27; per se. It&#x27;s more like a human vision system that samples light and hallucinates an appropriate image based on context. The image is constructed from the data more than it is reconstructed. And like human vision, it can be correct more often than not to be useful.
评论 #41280655 未加载
评论 #41278592 未加载
评论 #41287584 未加载
Dibby0539 months ago
This would be impressive if the examples weren&#x27;t taken from the same dataset (Laion-5B) that was used to train the Stable Diffusion model it&#x27;s using.
评论 #41279788 未加载
teruakohatu9 months ago
This is quite amazing that using a diffuser rather than a lens, then using a diffusion model can reconstruct an image so well.<p>The downside of this is that is heavily relies on the model to construct the image. Much like those colorisation models applied to old monochrome photos, the results will probably always look a little off based on the training data. I could imagine taking a photo of some weird art installation and the camera getting confused.<p>You can see examples of this when the model invented fabric texture on the fabric examples and converted solar panels to walls.
评论 #41277024 未加载
评论 #41280653 未加载
Thomashuet9 months ago
I don&#x27;t understand the use of a textual description. In which scenario do you not have enough space for a lens and yet have a textual description of the scene?
Snoozus9 months ago
It&#x27;s not as crazy as it seems, a pinhole camera doesn&#x27;t have any lenses either and works just fine. The hole size is a tradeoff between brightness and detail. This one has many holes and uses software to puzzle their images back together.
albert_e9 months ago
so this is like use of (in a different species) light sensitive patches of skin instead of the eye balls (lenses) that most animals on earth evolved ?<p>interesting.<p>even if this does not immediately replace traditional cameras and lenses... I am wondering if this can add a complementary set of capabilities to a traditional camera say next to a phone&#x27;s camera bump&#x2F;island&#x2F;cluster...so that we can drive some enhanced use cases<p>maybe store the wider context in raw format alongside the EXIF data ...so that future photo manipulation models can use that ambient data to do more realistic edits &#x2F; in painting &#x2F; out painting etc?<p>I am thinking this will benefit 3D photography and video graphics a lot if you can capture more of the ambient data, not strictly channeled through the lenses
mjburgess9 months ago
Does a camera without a lens make any physics sense? I cannot see how the scene geometry could be recoverable. Rays of light travelling from the scene arrive in all directions.<p>Intuitively, imagine moving your eye at every point along some square inch. Each position of the eye is a different image. Now all those images <i>overlap</i> on the sensor.<p>If you look at the images in the paper, everything except their most macro geometry and colour pallet is clearly generated -- since it changes depending on the prompt.<p>So at a guess, the lensless sensor gets this massive overlap of all possible photos at that location and so is able, at least, to capture minimal macro geometry and colour. This isn&#x27;t going to be a useful amount of information for almost any application.
评论 #41280710 未加载
评论 #41281032 未加载
xg159 months ago
Oh great, waiting for the first media piece where pictures from this &quot;camera&quot; are presented as evidence. (Or the inverse, where actual photographic evidence is disputed because who knows if the camera didn&#x27;t have AI stuff built in)
pvillano9 months ago
I wonder how it &quot;reacts&quot; to optical illusions? The ones we&#x27;re familiar with are optimized for probing the limits of the human visual system, but there might be some overlap
a1o9 months ago
Oh god, we are going to make lens a premium feature now aren&#x27;t we?
评论 #41276668 未加载
6gvONxR4sf7o9 months ago
Re: is this a camera or not, I recently realized that my fancy mirrorless camera is closer to this than i’d previously thought.<p>The sensor has a zillion pixels but each one only measures one color. for example, the pixel at index (145, 2832) might only measure green, while its neighbor at (145, 2833) only measures red. So we use models to fill in the blanks. We didn’t measure redness at (145, 2832) so we guess based on the redness nearby.<p>This kind of guessing is exactly what modern CV is so good at. So the line of what is a camera and what isn’t is a bit blurry to begin with.
评论 #41277207 未加载
评论 #41280869 未加载
valine9 months ago
I get the feeling that lens free cameras are the future. Obviously the results here are no where near good enough, but given the rapid improvement of diffusion models lately the trajectory seems clear.<p>Would love to lose the camera bump on the back of my phone.
评论 #41282137 未加载
评论 #41276859 未加载
ziofill9 months ago
+1 for the Thor labs candy box