Comment hidden because you are ignoring dag.(show it anyway)

I'm not sure I understand it, but he had me at "spacetime fusion technique"

Awesome.

Impressive!

Wow.

That is fascinating. Damn... When will it be available commercially? I'd download, I mean, buy it in a second.

Seriously though, that's kick ass.

lucky760 - What you are seeing here are results generated using some research quality code. There is no UI, no extensive error handling, and its not optimizied for speed because the code is usually written by one or two graduate students. The hope with such research projects is that some product team (e.g., Adobe After Effects) will read the publication and create a professional quality implementation of the ideas presented in the paper. But that usually takes a while to happen.

Impressive.

"Many effects shown in this paper are commonly created by folks in the visual effects industry using software packages such as Shake, BouJou, and AfterEffects. While the degree of automation in these packages has grown considerably in the past decade, they still rely heavily on manual effort (e.g., matting, rotoscoping, etc.) and specialized capture equipment (e.g., blue screen filming, steady cams, use of robotic arms to capture precise video paths, etc.); all are highly unappealing solutions to the amateur user."

Though similarly unappealing is prohibitively expensive software for the "amateur user." You undoubtable have more of a pulse on the industry that I. Do you suspect that we will be seeing significant cost reductions in this industry anytime soon? Sadly, even entry level video editing software of any decent quality has yet to manifest in the open source realm.

The porn industry would love this. Take a picture of some B-cup porn star and BOOM, double D's.

"Though similarly unappealing is prohibitively expensive software for the "amateur user." Do you suspect that we will be seeing significant cost reductions in this industry anytime soon?"

I think what is going to happen is that the visual effects industry is going to have to expand it's user base to include amateur users by creating software packages with the right pricing, UI, etc. Until recently there simply wasn't the automation technology nor the amateur auteur community of the size created by youtube (and other technologies) to realistically go after the amateur user market.

But obviously the market exits today and its ripe for harvesting. Just look at the difference between the quality of work produced by amateur photographers on Flicr and the quality of amateur videos on youtube. The difference is night and day. My feeling is that people really want to create aesthetically pleasing videos and edit them in interesting ways but they just don't have the right tools.

Hopefully the market forces will ensure that these tools are created and I for one can't wait for that to happen. Humans as a whole are a highly creative bunch. Its high time every visual storyteller out there had the tools to tell his/her story.

"Sadly, even entry level video editing software of any decent quality has yet to manifest in the open source realm"

Virtualdub is a freely available tool (and its open source) that I really like. But you are right, the current state of affairs leaves much to be desired.

"The porn industry would love this. Take a picture of some B-cup porn star and BOOM, double D's."

I didn't realize there were B-cup porn stars Unfortunately the B-to-DD cup transitions will probably have to be made by the cosmetic surgeons for now since the work being shown here can only handle videos of static scenes (i.e., videos shot with a moving camera but containing no moving objects in the scene). Efforts are being made to lift this restriction in future work.

Well its good to know the best minds are working on it right now.

They say the best innovations in computers were made because of war efforts, but if you ask me it was because of porn.

I am amazed by the level of sophistication in these tools. Go U Washington!

I fear that this work - created by graduate students - will be sold to profit the university for the exclusive use of some dreadful company like Salient Stills, and the product will be priced at $15,000 and will only be affordable by networks and intelligence agencies. Meanwhile, the grad students get nothing other than a degree and a job offer from the company that bought the technology. Please, please, please...open source this code. The end result will be far wider spread usage of the techniques. I can imagine a huge number of uses for these techniques, uses that will revolutionize video...but none of them will have any chance of ever coming to be if it is closed source and if the University sells this to one company. This is YOUR work, why should the Uni take it and sell it?

It looks to me like they could also change the path of the camera, as with the image stabilization example but to a greater extent. Awesome stuff.

It may be advanced, but where was the humor? Somebody throw a pie, for pete's sake!

In all seriousness, this went quite a bit over my head, but is very intrigueing. One of the, I think, simpler things, like removing the no parking sign is awesomely awesome.

Wow, I'm not the only one who thinks Salient Stills is probably the devil?

The porn industry helped bring cheap media storage. Storage that makes Youtube and the Sift possible.

Me, for some reason, liked the image stabilization tech the best. Need a steady cam/gantry shot and you have no budget? Duct tape a high megapixel still camera to your $150 digital movie cam and use this software...

*engineering

Adding video to channels (Engineering) - requested by oxdottir.