Removing obstructions and unwanted elements from digital images and photographs is a problem as old as mobile photography itself, but a research team from Google and MIT might just have the algorithm to solve our problems. Ever wanted a good photo of something beyond a chain-link fence? Or maybe have a good photo ruined by your own reflection on the window pane you’re shooting through?
Those issues might just be things of the past, if companies will look to the research of Michael Rubinstein, a research scientist at Google, and Tianfan Xue, a graduate student at MIT. Their work is based on the phenomenon called “motion parallax,”—this is when objects closer to us seem to move faster than those that are farther away. We’ve all noticed it before, but the researchers used this to mark the obstruction or offending reflection – since it is closer to the camera than the subject – and tells software which object to remove from the picture.
Rubinstein and Xue are aware that there have been some people who have attempted the same thing before them, but they believe that their algorithm can be more multi-purpose. Xue says that the process can also correct for a number of different kinds of obstructions on windows like raindrops or dirt. See the video below for more details on the algorithm.
Google is somewhat interested in the algorithm, and if developed, we can probably see this as a camera feature in the future. Note that most of the photos used in this research presentation was taken with an Android phone. The next step is to insert this feature into our smartphone camera apps so the masses can enjoy obstruction-free photography.
VIA: Technology Review
Goodbye digital watermarks!