We present a method for converting first-person videos, for example, captured with a helmet camera during activities such as rock climbing or bicycling, into hyperlapse videos: time-lapse videos with a smoothly moving camera.
At high speed-up rates, simple frame sub-sampling coupled with existing video stabilization methods does not work, because the erratic camera shake present in first-person videos is amplified by the speed-up. 3 researchers at Microsoft make a new type of video Time-lapse: Hyperlapse
I've always had this ironic feeling about MS hiring the best people they could get their hands on while never making anything that remotely reflects abilities of the kind. This innovation we have here however is of a level of complexity that... well... it is baffeling. It turns absolutely unwatchable garbage video that wouldn't do 10 views on youtube into something a movie studio would be proud of. It scores a 12 on the scale of -5 - 10 of my WTF 'o meter, right under remote viewing, the intergalactic federation, human stupidity and my own ideas of course. I should definetly get a new one.