This system, by students at the University of Washington, uses still photographs or one single frame of a video scene to automatically improve video. The improvements are amazing. Video might contain artifacts like overexposure and low-resolution imagery and this system takes cues from still images of the scene to bring almost the entire video up to photographic quality. Can’t wait for this to hit our local point and shoot – in maybe five years.
Comments 
Leave Comment
Commenting Options
Enter your personal information to the left, or sign in with your Facebook account by clicking the button below.
Alternatively, you can create an avatar that will appear whenever you leave a comment on a Gravatar-enabled blog.
- Actively Discussed Posts
- Friday Giveaway: Nixon Newton Watch
308 comments - Psystar's folly: 12 million clones per year by 2011
43 comments - You can install Chrome OS on your Dell Mini 10v right now
42 comments - Hong Kong Crunch: What's up in China?
26 comments - Square Enix boss claims console gaming on its last legs (and he's right)
22 comments
- There are no posts to display.
no way. i don’t believe it. that’s wicked crazy. especially the last one with the masking of the no parking sign. that doesn’t happen.
Frickin magic. That what I’m talkin about.
Ideally the high res photography camera and the video camera should be the same unit. Maybe taking a high res photo every second or two as a key frame while contiuously recording the video. Maybe even dynamically changing the rate of the High res camera dependent on the difference frame to frame.
Very cool. I wish I’d done it.
I think I also see that there technology is somewhat dependednt on side to side movement. I have made 3D movies out of 2D video using a similar technique.
Wow, that’s insane! The parking sign one is pretty hard to believe, I’d have to see the software do it with my own eyes.
All I can say is
http://www.AnythingNothing.com
I want it, I want it, I want it, NOW !
Once you are able to do the SFM (structure from montion) for the video and the picture and build the 3D, removing the poles and certain objects can actually be the easier part of the process (as long as you don’t have permanent occlusions).
As a complete system, this can have a major commercial impact on Hollywood and special effect for the movies. I particularly loved their (dynamic) texture mapping :-) Cut and paste 3D objects into each frame and… wow :-)
Cheers, E
That was without doubt the most amazing thing I have ever seen done with a video. Wow.
I wish this could get to the market sooner preferably now.
Daniel
This is revolutionary. CPU processing times are taking all the math out of filming.
Absolutely revolutionary. However when I see this, I can’t help but to think about the abuse and purposeful deceptions that can be disseminated to the public. I recognize that the same capability exists with still photos, but video is always more convincing. Imagine this software being used on video from Iraq that would otherwise show the carnage of war?
Believe me, as a photographer I am eagerly awaiting a production release of this technology so I can improve the quality of my work. I simply anticipate the inevitable realities of it’s capabilities.
The days of admissible video evidence are nearing their end.
How can anyone say the object removal part is hard to believe? You can do it in photoshop can you not? Yes you can, you take a few pictures from the side and basically fuse it into the one photo. All this does in the video is taking images from the angles that see behind the sign, removes the section where the sign is, then makes a finished image based on the the surrounding background and angled shots where you can see behind the sign. If you find it hard to believe, it’s only because you don’t know how it works.