Category Archives: Miscellaneous

3D capture & scanning, if you don’t mind looking like a tool

“Oh, is that True Love Waits conference?” my friend once snarkily asked as we drove past GPU conference attendees milling around downtown San Jose. “Is this Virgin-con?” Their dorktastic style comes to mind seeing demos for the helmet-mounted Wunder360.

Given that my trusty, if imperfect, Theta S 360º camera has gone MIA, I’m thinking about possible replacements. Having busted on the Wunder a bit, I’ll say I’m intrigued by the mapping possibilities. Given all it promises (especially relative to, say, the $499 Rylo camera), I’d worry that it’s oversold, especially at $159—but I guess we shall see.

The device promises:

  • Capturing 360 videos with in-camera stitching, no extra post-production software is needed;
  • Easy 3D scanning, enables the ability to create in 3D for everyone;
  • AI-powered smart tracking, locks on your favorite view;
  • Super smooth stabilization, say goodbye to shaky shots;
  • Compact, lightweight and portable, pop the S1 in your pocket;
  • With 100ft waterproof case, S1 works with you anywhere;


New MIT tech can visually isolate & remix musical instruments

I fully approve of this witchcraft. Now, to apply the “I forced a bot to watch 1,000 hours…” technique to my kids’ band recitals. 😉

VentureBeat writes,

The fully trained PixelPlayer system, given a video as the input, splits the accompanying audio and identifies the source of sound, and then calculates the volume of each pixel in the image and “spatially localizes” it — i.e., identifies regions in the clip that generate similar sound waves.




Demo: Synthesizing new views from your multi-lens phone images

You know the “I forced a bot to…” meme? Well, my colleagues Noah & team actually did it, forcing bots to watch real estate videos (which feature lots of stable, horizontal tracking shots) in order to synthesize animations between multiple independent images—say, the ones captured by a multi-lens phone:

We call this problem stereo magnification, and propose a learning framework that leverages a new layered representation that we call multiplane images (MPIs). Our method also uses a massive new data source for learning view extrapolation: online videos on YouTube.

Check out what it can enable:




The AI will see you now—specifically, right through your wall


The system works because those radio waves can penetrate objects like a wall, then bounce off a human body—which is mostly water, no friend to radio wave penetration—and travel back through the wall and to the device. “Now the challenge is: How do you interpret it?” Katabi says. That’s where the AI comes into play.

Now maybe we can get it running in your web browser, too. 🙂



WTF is with this dark pattern in iOS app trials?

Three of the free-to-download iOS apps I’ve tried in the last few days have led with a nasty trap for the less-than-vigilant: If you even want to try this app, agree up front to an expensive (like, orders of magnitude more than a usual app), ongoing subscription that quietly and perpetually renews until you figure out how to stop it.

This upends the normal trial relationship of “If & only if you like this offering enough to buy/subscribe, take action to do so; otherwise you’re off the hook.” Screw that: these apps are getting summarily shitcanned.

Meanwhile I’m amazed that Apple allows this practice to continue.


Machine learning in trees: Cupertino students build a Smart Wildfire Sensor

Kids these days: Two high school students used TensorFlow, Google’s open-source machine learning tool, to build a Smart Wildfire Sensor, 

Aditya Shah and Sanjana Shah, two friends and high school students from Cupertino, California… built a Smart Wildfire Sensor, which can help predict where wildfires are most likely to occur in a given area so firefighters are better prepared to stop them.



“Deep Video Portraits” can put words in your mouth

Part of me says, “What great new tools for expressive video editing!”

The other part says, “This will not end well…”

[W]e are the first to transfer the full 3D head position, head rotation, face expression, eye gaze, and eye blinking from a source actor to a portrait video of a target actor… [W]e can reenact the full head using interactive user-controlled editing, and realize high-fidelity visual dubbing.


[YouTube] [Via Jeremy Cowles]

The joyously absurd Rube Goldberg machines of Joseph Herscher

I’m getting way too big a kick out of the work of kinetic artist & toymaker Joseph Herscher:

Khoi Vinh writes, in an inventory worth of Stefon (“this place has everything…”),

Herscher’s pièce de résistance may be “The Cake Server,” shown above: a gorgeous monstrosity that brings together melting butter, a glass of juice that pours its contents into itself, a baby using a smartphone and much more to serve a slice of upside-down cake to a plate in its God-intended manner of delivery. It’s a marvel to behold.



[YouTube 1 & 2]