Google and MIT showcase photography algorithm that eliminates reflections

If you’ve ever watched an episode of Law & Order or CSI, you have probably noticed some outlandish forensic work involving photographs being used to aid the show’s team of detectives in furthering their investigation. Sometimes it’s clearing up pixelations or using a minute reflection in a window to read some perp’s name tag, which are all things that’s mostly cooked up in fantasy.

Until now…

In a partnership between Google and the Massachusetts Institute of Technology (MIT), breakthrough work in photography algorithms has allowed researchers to eliminate, or mostly eliminate, occlusions and reflections from photographs. This is the part that’s great for consumers taking photographs at zoos and out of windows.

In order to do this, the image separates the background from the occlusion or reflection. This means that by the camera isolating the obstruction, it’s able to render that as a completely separate image with amazing clarity. This is the part good for law enforcement.

To get a better understanding of how all of this works, MIT researcher Tianfan Xue has uploaded an awesome YouTube video demonstrating the algorithm, which you can view below.

Click here to view the embedded video.

Xue will be joined by his fellow research partners in presenting a paper over their work at Siggraph 2015 this month.

Source: Google

Come comment on this article: Google and MIT showcase photography algorithm that eliminates reflections