TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

AI camera with no lens

290 点作者 anitakirkovska大约 2 年前

24 条评论

belugacat大约 2 年前
I can&#x27;t find it now, but there was a prototype someone made in the 2000s of a camera that, when you pressed the shutter, would fetch an image on Flickr that most closely matched your GPS coordinates + time of day, acting in a similar way as a &quot;crowdsourced camera with no lens&quot;.<p>Fun to the see a modern reincarnation of that idea.<p>(While digging around to find the above, I did find yet another camera project that does the opposite: <i>&quot;Matt Richardson&#x27;s &quot;Descriptive Camera&quot; sends your pictures to Amazon&#x27;s Mechanical Turk and jobs out the task of writing a brief description of each image, then outputs the text on a thermal printer. It&#x27;s a camera that captures descriptions, not pictures.&quot;</i> (<a href="https:&#x2F;&#x2F;boingboing.net&#x2F;2012&#x2F;04&#x2F;25&#x2F;descriptive-camera-prints-out.html" rel="nofollow">https:&#x2F;&#x2F;boingboing.net&#x2F;2012&#x2F;04&#x2F;25&#x2F;descriptive-camera-prints-...</a>)
评论 #36144297 未加载
评论 #36140185 未加载
评论 #36140269 未加载
评论 #36166006 未加载
dmbche大约 2 年前
If anything, this is a neat art piece asking &quot;Why do people take pictures?&quot;.<p>I&#x27;ve found often that the gut feel that makes you take the shot doesn&#x27;t necessarily know when you have the right composition or what the subject of the shot actually is, you just hit he shutter, you know that this was the shot.<p>Then, when looking at the shots, you have all the time and the world to analyse and find meaning and beauty in this sliver of an instant.<p>By replacing this by a random seed, a 20 word prompt and gps localisation, I doubt that anyone would have a personal connection to the image, or to the instant it was taken. It become a &quot;clean&quot;, &quot;sanitized&quot; image, that&#x27;s only esthetic (or arguably memetic depending on your prompt), and is wholly separate from the person that took it.<p>You also lose all of the information that you can not consciously perceive while taking the shot&#x2F;writing the prompt, since you filter what you see through the lens of language, and then back into visual.<p>It&#x27;s neat !
评论 #36143014 未加载
评论 #36144006 未加载
Solvency大约 2 年前
Idea: AI police sketch artist.<p>3D printed case. Resting on a table. Witness describes the suspect. 10 seconds later it prints out an AI generated ink sketch.<p>I mean why not? Sell it through to every police station in your state. You can even put a cute little police badge emblem on the case.
评论 #36140302 未加载
评论 #36141585 未加载
评论 #36149325 未加载
评论 #36141310 未加载
评论 #36140084 未加载
评论 #36140179 未加载
评论 #36146649 未加载
评论 #36140032 未加载
leononame大约 2 年前
Cute and fun idea, but it&#x27;d be nice if it could take better indoor photos.<p>Jokes aside, I think this demonstrates that AI generation isn&#x27;t too great if you have something very specific in mind, at least it looks like the generated picture deviates from the real one, though it&#x27;s impressive that it&#x27;s still so extremely similar.<p>The virtual one doesn&#x27;t load for me unfortunately
评论 #36139962 未加载
boo-ga-ga大约 2 年前
Really love the concept. There also was a project with instant camera that draws cartoons based on what is captured: <a href="https:&#x2F;&#x2F;web.archive.org&#x2F;web&#x2F;20181003104348&#x2F;https:&#x2F;&#x2F;danmacnish.com&#x2F;2018&#x2F;07&#x2F;01&#x2F;draw-this&#x2F;" rel="nofollow">https:&#x2F;&#x2F;web.archive.org&#x2F;web&#x2F;20181003104348&#x2F;https:&#x2F;&#x2F;danmacnis...</a>
manx大约 2 年前
There was also a camera in 2012, which sent its picture to mechanical turk to ask for a description: <a href="https:&#x2F;&#x2F;hackaday.com&#x2F;2012&#x2F;04&#x2F;25&#x2F;a-camera-that-describes-a-picture-for-you&#x2F;" rel="nofollow">https:&#x2F;&#x2F;hackaday.com&#x2F;2012&#x2F;04&#x2F;25&#x2F;a-camera-that-describes-a-pi...</a>
评论 #36142341 未加载
ayoreis大约 2 年前
I thought this was going to be a more advanced version of this: <a href="https:&#x2F;&#x2F;petapixel.com&#x2F;2022&#x2F;05&#x2F;04&#x2F;newly-developed-camera-can-take-in-focus-photos-without-a-lens&#x2F;" rel="nofollow">https:&#x2F;&#x2F;petapixel.com&#x2F;2022&#x2F;05&#x2F;04&#x2F;newly-developed-camera-can-...</a>
dan-g大约 2 年前
Original source: <a href="https:&#x2F;&#x2F;twitter.com&#x2F;BjoernKarmann&#x2F;status&#x2F;1663496103998750721" rel="nofollow">https:&#x2F;&#x2F;twitter.com&#x2F;BjoernKarmann&#x2F;status&#x2F;1663496103998750721</a>
评论 #36144929 未加载
TheDudeMan大约 2 年前
At first, I thought, &quot;Cool, a sensor but no lens.&quot; Nope, no sensor.
评论 #36158813 未加载
nonethewiser大约 2 年前
I see no reason to consider this a camera. Its just an image generator.
评论 #36141306 未加载
评论 #36141223 未加载
评论 #36141125 未加载
amelius大约 2 年前
I guess you can take photos of the world and turn everything into a giant NeRF.<p><a href="https:&#x2F;&#x2F;www.plainconcepts.com&#x2F;nerf-3d&#x2F;" rel="nofollow">https:&#x2F;&#x2F;www.plainconcepts.com&#x2F;nerf-3d&#x2F;</a>
vngzs大约 2 年前
This should link to the original post, not link to a blog post about a Twitter thread that contains a link to the original post.<p><a href="https:&#x2F;&#x2F;bjoernkarmann.dk&#x2F;project&#x2F;paragraphica" rel="nofollow">https:&#x2F;&#x2F;bjoernkarmann.dk&#x2F;project&#x2F;paragraphica</a><p>The author&#x27;s site is currently struggling to load, so here is an archive link:<p><a href="https:&#x2F;&#x2F;archive.is&#x2F;mvfq2" rel="nofollow">https:&#x2F;&#x2F;archive.is&#x2F;mvfq2</a>
icepat大约 2 年前
Interesting, but (to me at least) the point of photography will always be the technical process behind it, and capturing a _specific_ aspect of the environment. This will turn out fine generic images, but if you want to capture more, that&#x27;s not something you can just synthesise.<p>This isn&#x27;t really a camera, it&#x27;s a GPS hooked up to Stable Diffusion.
yowzadave大约 2 年前
Unfortunately the major weakness of this camera is the one thing I actually use my camera for: photos of my kids (and other people I know). But for stuff like landmarks, I never even bother taking photos of them—I’ll never take a photo of Delicate Arch that is as good as one million others I can find by doing an image search.
peddamat大约 2 年前
So it&#x27;s basically a Samsung.
bafe大约 2 年前
The title is really misleading. I was expecting a camera without lens, using a type of light sensitive array and reconstructing the focused image using AI
dist-epoch大约 2 年前
Oh dear, see through cameras are about to become a thing.
PinkMilkshake大约 2 年前
Very cool idea. I wonder if the design is a reference to the star-nosed mole. Appropriate for a blind camera!
评论 #36144831 未加载
m463大约 2 年前
this kind of technology is just what we need. It will always take a photo of the past, with more youthful self, hallucinated.
m3kw9大约 2 年前
Even if you bring that back to the 80s they’d think you are doing witch craft if you showed this to anyone
评论 #36145410 未加载
random3大约 2 年前
use-case, use with AR glasses to live enhance current location
yieldcrv大约 2 年前
exhibit a of how prompt engineering wont be a valued salaried skill, it is a valuable skill to create a revenue stream from
sgok大约 2 年前
it&#x27;s like magic. incredible.
teacpde大约 2 年前
source site getting HN hug?