13 comments

  • functionmouse6 hours ago
    I feel like the primary use case for such a technology is manipulating and profiling people over video chat, maybe even autonomously. Hiring managers, HR, landlords, and police are obvious customers.<p>The response I anticipate will be &quot;But this will help doctors over telehealth and stuff!&quot; - Please see <a href="https:&#x2F;&#x2F;calebhearth.com&#x2F;dont-get-distracted" rel="nofollow">https:&#x2F;&#x2F;calebhearth.com&#x2F;dont-get-distracted</a>
    • nearbuy5 hours ago
      This tech (detecting pulse from regular video) has been around almost 20 years now, and this doesn&#x27;t seemed to have happened yet.<p>You see this type of thing in spy movies, but I&#x27;m not sure it&#x27;s that useful in real life. You&#x27;re basically taking one piece of data a polygraph uses, but without the most important component (skin conductance). Polygraph accuracy isn&#x27;t that great to begin with. You can profile and manipulate people more effectively based on their reactions and behaviour, and their pulse will be much harder to interpret.
      • ranger_danger4 hours ago
        I don&#x27;t think this tech has actually been used in practice for that long, if at all. It was only first demonstrated in 2012 at SIGGRAPH.<p>Can you cite any commercially available uses of such tech?
        • nearbuy3 hours ago
          I don&#x27;t know any commercial uses of such tech <i>today</i>. I&#x27;m not saying they don&#x27;t exist. I just don&#x27;t know of them.<p>I had said I don&#x27;t think it&#x27;s very useful for &quot;manipulating and profiling people over video chat&quot;, so I wouldn&#x27;t really expect there to be a commercial product for that. Probably it&#x27;s used in fitness or heart rate monitoring apps for people that don&#x27;t have a fitness tracker device and prefer not to manually count their pulse.<p>Here is the tech demonstrated in 2007: <a href="https:&#x2F;&#x2F;pubmed.ncbi.nlm.nih.gov&#x2F;17074525&#x2F;" rel="nofollow">https:&#x2F;&#x2F;pubmed.ncbi.nlm.nih.gov&#x2F;17074525&#x2F;</a><p>The core algorithm is really simple. You find a patch of skin. Take the average color of the pixels in that patch. The color will become more reddish each pulse. Do an FFT and take the strongest peak in the plausible heart rate range. You could prototype this in a few hundred lines of python.<p>If this were useful for police or hiring managers, someone could have use the tech to make an app for them within the past 19 years.<p>Of course, companies have a history of trying to market a lot of BS metrics (e.g. graphology, MBTI) to hiring managers, so I wouldn&#x27;t be that surprised to see a company claim they can predict employee success using pulse. Whether it works is another story.
          • ddalex2 hours ago
            &gt; You could prototype this in a few hundred lines of python.<p>You mean Claude can one-shot this.
        • swiftcoder59 minutes ago
          I don&#x27;t it&#x27;s ever been practical to ship in a product? You need ~20 seconds of data to stabilise the reading, and any large motion ruins it - even though Microsoft Research demonstrated a Kinect could detect heartrate in a lab setting, it wasn&#x27;t viable to ship in a fitness game.
        • numpad04 hours ago
          (2008)<p>1: <a href="https:&#x2F;&#x2F;pmc.ncbi.nlm.nih.gov&#x2F;articles&#x2F;PMC2717852&#x2F;" rel="nofollow">https:&#x2F;&#x2F;pmc.ncbi.nlm.nih.gov&#x2F;articles&#x2F;PMC2717852&#x2F;</a>
    • metalcrow6 hours ago
      Can you explain how <a href="https:&#x2F;&#x2F;calebhearth.com&#x2F;dont-get-distracted" rel="nofollow">https:&#x2F;&#x2F;calebhearth.com&#x2F;dont-get-distracted</a> applies to the potential response you described? I don&#x27;t get it.
      • croes6 hours ago
        They will weaponize it.
      • godelski5 hours ago
        Don&#x27;t get distracted, sit down and read it in full.<p>Don&#x27;t get distracted, think about what he wrote.<p>If you still don&#x27;t get it, take a step back. Think. Process. Then take a break and read it again tomorrow.<p>Slow down. Don&#x27;t get distracted. You don&#x27;t need to respond so fast. Take your time. There is no rush. There is no shortcut. Read it in full and you&#x27;ll understand this comment says much more.
        • _Microft1 hour ago
          Here‘s also an advice: if you want someone to listen, try not to come across like you just did.
          • godelski48 minutes ago
            Your sibling said something similar, my response is identical<p><a href="https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=47294347">https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=47294347</a>
            • _Microft41 minutes ago
              Sure, I read that but your comment still sounds like it does. You&#x27;re doing yourself no service.
          • peyton1 hour ago
            We’re just BSing on the internet. No need to tone police.
            • _Microft1 hour ago
              <i>That</i> internet is elsewhere.
        • collingreen4 hours ago
          This feels overly patronizing
          • godelski4 hours ago
            Probably because I repeated &quot;don&#x27;t get distracted&quot;. But if you read the article then I think it&#x27;ll take on a different context, as I&#x27;m mimicking the author, including their short paragraph style.
            • user27221 hour ago
              I get really annoyed at those articles which advocate the developer to sacrifice themselves towards a better future.<p>Companies externalize costs. I refuse to be the one, as an individual, with the burden of fixing society ills to my own detriment.<p>Tell me to get into politics, join an association, whatever. Now, as an individual, lose money for morals? No thank you. I may, and probably will, do it -- but don&#x27;t <i>expect</i> I do it. I have no business, in a society with less and less public services, to harm myself and my family for refusing to do well paying jobs.<p>I will externalise those costs as much as possible. I will bring awareness. I will write letters. But don&#x27;t ask me to leave a well paying job -- that&#x27;s someone else job to fix.
              • godelski42 minutes ago
                <p><pre><code> &gt; as an individual </code></pre> But that&#x27;s the problem. Your logic applies to everyone in an organization (a business, a family, a country, and so on). The organizations actions are not the result of any single actors decisions, even if weight isn&#x27;t equal. The decisions of an organization are made of the decisions of the collective. The agglomeration of them. And that&#x27;s why everyone&#x27;s decisions matter. Because you don&#x27;t know when your actions have more weight than when they have less.<p>We&#x27;re all in this together. One way or another, your actions affect others. Your actions aren&#x27;t in isolation. Conversely this is true for others, and I suspect you would rather others treat you well, right? So which feedback loop do you want you contribute to? That&#x27;s the only question there is
              • _Microft47 minutes ago
                ´&quot;That&#x27;s not my department&quot;, says Wernher von Braun.´
  • _Microft1 hour ago
    You as reader might also be interested in this: <a href="https:&#x2F;&#x2F;hbenbel.github.io&#x2F;blog&#x2F;evm&#x2F;" rel="nofollow">https:&#x2F;&#x2F;hbenbel.github.io&#x2F;blog&#x2F;evm&#x2F;</a>
  • twodave6 hours ago
    It’s not very accurate. Maybe because of the camera fidelity. It was about 10bpm lower than actual for me. Seems to operate off of subtle motions caused by pulse. It was even worse at detecting breathing.
    • brynnbee3 hours ago
      It showed my BPM at nearly half of what it really was
  • jmusall5 hours ago
    May be related: Explanation of motion and color amplification in video by Steve Mould <a href="https:&#x2F;&#x2F;youtube.com&#x2F;watch?v=rEoc0YoALt0" rel="nofollow">https:&#x2F;&#x2F;youtube.com&#x2F;watch?v=rEoc0YoALt0</a><p>He even shows pulse detection (around 8:30).
  • smusamashah3 hours ago
    I made my own heart rate app (using gemini at first and then Claude for lots of further edits). <a href="https:&#x2F;&#x2F;xosh.org&#x2F;heart-rate&#x2F;" rel="nofollow">https:&#x2F;&#x2F;xosh.org&#x2F;heart-rate&#x2F;</a> <a href="https:&#x2F;&#x2F;github.com&#x2F;SMUsamaShah&#x2F;heart-rate" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;SMUsamaShah&#x2F;heart-rate</a> it&#x27;s all offline. UI needs more work but me and my wife are the only users therefore it doesn&#x27;t matter that much.<p>At one point I added the same EVM based heart rate detection but that requires sitting very still. I use it on my phone mainly therefore the common finger method is easiest one to use.
  • sxp5 hours ago
    How does it work? Is it <a href="https:&#x2F;&#x2F;people.csail.mit.edu&#x2F;mrub&#x2F;evm&#x2F;" rel="nofollow">https:&#x2F;&#x2F;people.csail.mit.edu&#x2F;mrub&#x2F;evm&#x2F;</a>? I see the FAQ about VitalLens, but I couldn&#x27;t find technical details.<p>It&#x27;s super cool. Thanks for sharing. I want to build a biofeedback app for meditation and this looks like a good platform to use.
    • serious_angel4 hours ago
      Please consider checking out comments: <a href="https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=47293662">https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=47293662</a>
      • whilenot-dev1 hour ago
        What do you mean? Eulerian-Video-Magnification could be (and probably is) the underlying algorithm in the VitalLens API.
  • amagasaki7 hours ago
    Could be interesting, but allowing the webcam crashes my browser.Repeatedly<p>macOS 15.7.1 (24G231) Brave 1.87.186 (Official Build) (arm64) Chromium: 145.0.7632.45
    • serious_angel6 hours ago
      In the minified source code, we may see, it uses:<p>``` try { const l=await navigator.mediaDevices.getUserMedia({audio:!1,video:{facingMode:&quot;user&quot;}}); &#x2F;* ... *&#x2F; } catch { this.showError(&quot;Could not access webcam. Please check permissions.&quot;) } ```<p>There are alternatives to verify mediaDevices support as <a href="https:&#x2F;&#x2F;addpipe.com&#x2F;getusermedia-examples&#x2F;" rel="nofollow">https:&#x2F;&#x2F;addpipe.com&#x2F;getusermedia-examples&#x2F;</a>
      • echoangle5 hours ago
        But even a missing check shouldn’t be able to crash the browser.
  • hluska6 hours ago
    I would prefer some kind of privacy statement or even some kind of explanation about what is going on before I just randomly turn my webcam on. This might be great and I’m proud of you for launching but I don’t do things like that. Heck, videos can make a person’s heart race - I had my first attack at 39 and that’s a hell of a lot of risk.
    • serious_angel6 hours ago
      I haven&#x27;t dug deeper due to time availability, but for the same sake of privacy, I&#x27;ve found:<p>1. `&#x2F;api&#x2F;event` endpoint mentioned in the `&#x2F;stats&#x2F;script.js` file;<p>2. There&#x27;s `&#x2F;parties&#x2F;lobby&#x2F;main&#x2F;telemetry` in a minified JavaScript chunk asset;<p>3. There&#x27;s VitalLens mentioned, and there&#x27;s an error string in the same asset: &quot;A valid API key or proxy URL is required to use VitalLens. If you signed up recently, please try again in a minute to allow your API key to become active. Otherwise, head to <a href="https:&#x2F;&#x2F;www.rouast.com&#x2F;api" rel="nofollow">https:&#x2F;&#x2F;www.rouast.com&#x2F;api</a> to get a free API key.&quot;
      • hluska6 hours ago
        This is really kind of you - I appreciate this. Thank you for taking that time!
  • trothamel5 hours ago
    There&#x27;s a version of this built into the Google Fit application for Android.
  • gumboshoes7 hours ago
    Worked for me on Android. Love the simplicity.
  • abakker6 hours ago
    It registered in the low 40s for me, while my watch was saying 72-75. I guess YMMV
  • Levelupcarrent1 hour ago
    [dead]