Slashdot Mirror
Turn Real Life Into A Cartoon
Saige writes "Ever wanted to see yourself in a cartoon? Before now, there were means to turn a single image into something cartoon-like, but some folks at Microsoft Research have come up with a method to turn a video into an animated cartoon. It's not up to doing it fully automated, as you have to hand-mark various parts of the video every 10 to 15 frames, but the video of the results is quite impressive."
That's it, good night folks, I've seen it all.
"However, even the 300 frame video of the girl swinging on the money bars only needs a keyframe every 10 or 15 frames."
I just hope they don't make it part of Wordart or something.
US Democracy:The best person for the job (among These pre-selected choices...)
I always knew Microsoft was a Mickey Mouse corporation.
Nothing disturbs me more than blind loyalism towards some unrealistic and over-idealistic notion of one's nationality.
Bill's kid comes home crying. Seems his little schoolchum's dad Steve has a hip movie studio that makes way cool animated features. Why can't you do that, Dad? I want an animation studio! I want it right now! So Daddy Bill picks up the phone and commands that Megacorp also begin work on animation. Unfortunately, Megacorp's work ends up looking a lot like old Clutch Cargo episodes. Bill's kid cries himself to sleep.
If Slashdot were chemistry it would look like this:Cadaverine
My life is already quite cartoonishly silly. The only way software could make it more so would be to automatically drop an anvil on me as I fell off a cliff like Wylie Coyote.
Similar to Waking Life, one of my all-time favorite movies. On the dvd, there's a 20 minute segment explaining the technology behind it...very labor intensive, as every curve ultimately still had to be hand-done.
And now for something completely different...a man with three buttocks.
I tinker occasionally with animation and despite all the technology we have today, if you are a 2D/cel animator it's still an extremly slow process. But fun.
no, but it would have made Waking Life and Bakshi's Lord of the Rings a lot easier.
I'll be most impressed when they have a Cartoon Physics Engine.
I used to work at a place where some sort of technique was applied to turn the work environment into a Dilbert episode.
If you're running at a good clip per second, that's several frames per second that you're giving it animation information. As the microsoft researcher says, it's interpolating between keyframes, smoothing for trajectory. It's probably also taking averages of color inbetween the frames, and running it through a natural media highlight algorithim. Think those oldfangled "morph" programs mixed with a photoshop filter.
It should be doing some edge detection for the inbetween frames, but it probably isn't. I hate to say this, but this is a simple application of known and existing technologies. Nifty for the guys that made it, but not exactly groundbreaking.
The ______ Agenda
My first thought-- oh, great. Put me out of work.
But then I came to my senses. Of course this kind of thing would never replace traditional animation. After all, you'd still have to have actors enact the scenes to be animated, the backgrounds would have to be set up or altered, etc. Setting up a shoot of a scene to be animated could end up being more of a PITA than just animating it to begin with. Though the end result could be a cool rotoscope/Waking Life effect, it's not a "cheat" to get an animated feature without the tedious work of animating.
this? http://students.washington.edu/juew/
This is basically a way of partially automating the process of rotoscoping, which goes back to the 1930s. It's not generally used because the resulting animation looks choppier and less cartoon-like; it's the reason why Ralph Bakshi's later films (Lord of the Rings, Cool World, American Pop) generally are considered not to look as good as his previous films: they were almost entirely rotoscoped.
Bít, zabít, jen proto, ze su liska!
Isn't it interesting how throughout the last several years we've been researching and coding like hell to take cartoon(ish) characters and make them look as realistic as possible? Look at the work that went into transforming an artist's sketches of Dr. Aki Ross et al into the very real looking characters of Final Fantasy.
Now we're researching and coding like hell to go back the other way.
I'm sure there's a Microsoft joke in there somewhere :)
You've got an easy breezy wind at your back...most of the time.
Anybody remember this guy?
This is one of the pioneers in computer graphics for a long time. You should remember him for his radiosity papers:
Cohen, M. F. and Greenberg, D. P., "The Hemi-Cube: A Radiosity Solution for Complex Environments", Computer Graphics, vol. 19, no. 3, pp 31-40, 1985.
Cohen, M. F., Chen, S. E., Wallace, J. R., and Greenberg, D. P., "A Progressive Refinement Approach to Fast Radiosity Image Generation", Computer Graphics, vol. 22, no. 4, pp 75-84, 1988.
And his book.
He even received SIGGRAPH award for his work
Just a thought. I've played with Photoshop/Paintshop Pro and various standard filters can turn individual photos into an artistic rendering eg. Brushstrokes or Charcoal drawing. What's to stop someone from writing software that will extract each image from a video, apply the filter and then re-encode to video? Has this already been done elsewhere?
As an aside I love the effect on pets using the charcoal filters drawing filters. The fur translates surprisingly well.
These posts express my own personal views, not those of my employer
Some time ago, Microsoft purchased a company called SoftImage. Turned out to be a good investment in 3D development and film compositing with a product called the DS.
Meanwhile, in Tewksbury, the Avid Media Composer which ran only on the Apple Macintosh platform was ported to Windows when Microsoft made some investments in Avid. About that time Apple (unwisely) discontinued their six PCI-Slot Macintosh..
When Avid noted that their product was dead-ended because its code basis assumed a raster that was limited to NTSC and PAL television format, they purchased SoftImage's DS in order to be able to easily produce software that will do film and high definition video.
Microsoft doesn't make investments for nothing. I believe I can do something very close to what Microsoft is doing for Mini-DV video on any format of video or film with the Avid DS -- though for a lot more money (something like $120K USD). I would not be surprised if they got the technology from that very old investment.
As a creative person though, I have to say I don't like the fact that the DS-Nitris will probably never run on a Macintosh. We have problems with ours that are related mostly to two issues: Operator screw-ups (expected) and Microsoft Windows XP Professional limitations, many of which do not exist in Apple's current versions of Unix.
Gods don't kill people, people with gods kill people.
Like any large company, there are many different departments handling many different things.
Research is but one of those departments. And why deny them the ability to do further research? In the end, with what they've learned doing research it can only help their products that are already out in the market.
So basically MS is taking credit for work largely done by three Asian graduate students? Kind of like three Ph.D. students at Harvard finding a cure for AIDS, and then Harvard claiming it's their discovery.
Cohen's colleagues get zero name recognition in the MS article. Kind of awkward don't you think? It comes off as if the other workers' contributions are insignificant.
The parent is still very informative. We wouldn't have even known about the other contributors if it weren't for him.
And anyone who has worked under a big-name advisor on a project knows they have a tendency to take credit for more than they actually did, especially when foreign students are involved.
Not again, thanks! I already found it here.
Regards, Martin
That's called Rotoscoping, and it's been around since before the original Lord of the Rings movie by Ralph Bakashi (1978).
That's not what the article was about, really, if you read it. Rotoscoping is modifying each frame individually, in a manner similar to how you do a cartoon.
If you RTFA (fat chance, I know), the article addresses this: "In addition, current techniques to turn videos into cartoons are very labor intensive; the artist has to render each frame by hand. And it still doesn't solve the 'jumping' problem.".