Copy
Hi Fives readers! (pun?)
Over the last couple weeks I've received amazing feedback on Fives and how it can be improved. You'll see some small changes over the next few weeks reflecting that feedback (of course, only the feedback I agree with). Keep it coming, love to hear what you think!

This week's been a little crazy so I'm a day late. I'll make it up to you with an extra-awesome edition of Fives. 
Share
Tweet
Forward
+1
Share
1. Light in Motion
There's been some work over the last few years (a lot in my group at the Media Lab, actually) in which researchers have found ways to film the propagation of light through a scene using cameras that can capture images at a trillion frames per second. This is yet another iteration of that. We all know mirrors reflect light, but have you ever seen it actually bounce off? Now we can. Researchers have have used a streak camera (high-speed 1-dimensional imaging system) and some clever optics and reconstructions to image light as it bounces off of a mirror. You can actually see the packet of photons come off at exactly the angle you can expect. With these sorts of techniques, we may be able to better understand and design high-speed communications (fiber optics) or even film the progress of high-speed chemical reactions. [Video + More]
2. Lensless Cameras
Though lensless cameras have existed for some time (see: pinhole camera), a company called Rambus has published a video demonstrating a particularly cool one. Their clever use of a diffraction grating pattern (holes/shapes in the optics) turns incoming light into a spiral pattern detected the camera's sensor. Then, with some interesting reconstruction algorithms, they can rebuild the encoded image. Why is this so great? Well, the process of making these diffraction gratings is far cheaper and far less complex than current lens-based cameras. It would also make camera's much much smaller. Though the extra computation definitely isn't a huge plus. [More]
3. Automated Material Recognition
This is particularly amazing. Researchers have used a TON of data, some smart preprocessing, and a drop of machine learning (Convolution Neural Network) to automatically identify and segment different materials in an image. That is, purely through some training on textures, the algorithm can recognize that a couch material is different from the floor or plant material. Check out the results in the paper, you may even notice it realizes things you yourself don't! Why is this cool? One day we might use this to build more intelligent algorithms for photo-editing or use it for some even smarter automated captioning on Google Search.  [More]
4. High Definition Eye Capture
See that picture above? It's a 3D model. Just watch the video first, it's 90% of the description. Did you do it? Isn't that an impressive result? With just a few cameras and some good lighting Disney researchers can capture your eye in extremely high detail and simulate how it might even respond to light cast on it. If Disney's working on it, graphics are about to get a lot more impressive in your next Pixar film. Especially now that they can cheat and use real eye data. [More]
5. YotaPhone's Half E-ink Smartphone
I love shining some light on the lesser known hardware companies out there. This news is a couple weeks old, but I felt it was worth sharing with you all. A Russian company has debuted a phone called the "YotaPhone 2" This phone has a regular screen on one side and a slick, inset e-ink display on the other. Why is that useful? Always-on notifications that don't drain your battery, Google maps n the back so you can navigate for days, or even a persistent shopping list while you're getting groceries; take your pick. Tons of interesting use-cases and an admirable departure from current smartphone "biggovations". [More]
Fives is written by Rohan Puri