Photosynth Turns Photos Into Immersive Environments

This week, one of Microsoft’s impressive lab products Photosynth was finally opened for personal use after 2 years of being only a technology demo.  Honestly even after two years I never expected it for public use given it’s seeming unheard of photo analysis technology, but this weekend I was able to dive in and start using Photosynth for myself.

Photosynth takes batches of photos (20+ photos at bare minimum) and analyzes them for overlapping content and views, and then builds them into a 3D environment where you can click into photos and see the objects within from different views, as well as to continue to move around environments as if all of the pictures are one. 

It’s hard to really comprehend until you start to use it, so I’ve embedded a photosynth above that I built of Mirror Lake in the Wallowa Mountain Range in Oregon (pictures and more from my recent backpacking trip).  When you upload pictures, Microsoft Photosynth tells you how “synthy” they are, which basically reflects how well your photos really connect with each other to build an environment.  You can see my Wallowa photosynth is 59% synthy, so it’s actually a pretty good collection of 44 pictures, but they split out into a few seperate groups and a few pictures can’t be connected.

One thing to note – Microsoft Photosynth is very buggy.  You may have to install it numerous times, and often when I return to the site it cannot detect  that I’ve already installed the plug-ins.  At any given time, either Firefox or IE doesn’t allow me to actually view the photosynths.  Hopefully over time the site will become much more stable, which is necessary for it to scale.

I’ve also built one from pictures I took of the Taj Mahal on my trip to India, see both of my photosynths here.