This is a really cool technique that posing interesting questions about how 3D resources in the future will exist.
Will marketing companies release models and textures like we release different image formats now? Are there grest uses for quickly making video or 3D from a photo. IDK, really cool though.
The supplement mentions this about backside texturing:
```
Texturing the back of the body is more difficult, as we have no direct observation of it. One approach is to sim- ply mirror the front texture onto the back. This mirrored- texturing produces reasonable results in some cases (e.g., arms), but undesirable results in others (face appears on the back of the head). To address this problem, we allow the user to choose between mirrored texturing or label-driven texture synthesis – “texture-by-numbers” – on a part-by- part basis.
```
Looks super cool. I wonder if multiple subjects in the same photo could be animated. I sorta expected that from the picture with the Beatles ( assuming that gang of 4 Brittish looking guys was them )
Very impressive. Not to be a cynic, but how long before someone starts turning people inappropriately into frog marching saluting Nazis?! Or giving them an unfortunate Fortnite "bounce".
I haven't read the paper, but am curious how they get the face profile right (given the information is absent in the 2D source). Perhaps they set that manually or some kind of average face, but if it's off it would be quite noticeable.
That artwork is basically saying kids should be flying kites, rather than being bombed by drones while they cry in terror.
I find playing around with that tasteless, to put it mildly and charitably, and I said nothing about it, after seeing it and puking into my mouth a little -- if walking on egg shells, and calling it tasteless to be generous, is punished anway, I may as well be honest, and I'm still biting my tongue on just how tacky CS meeting art is 99% of the time, that's tasteless too but in the actual meaning of the word, just unskilled, garish, time and time again demonstrating that developed art comes from developed persons, not from developed tools -- right after the story was posted and there were no comments. I just shook my head basically, kept it to myself. But when I later saw the question of how long it'll take before someone does something inappropriate with it, I just had to point it out.
What guideline was I violating? How was my comment not a contribution? If you disagree with it being tasteless, if I overlooked or misrepresented something, where are the words?
FWIW I think you’re completely correct. The use of that image was near as tone deaf as I can imagine. It almost seems intentional.
With a very charitable interpretation it could be said that it’s a warning, a harbinger of the callousnness with which this tech will be used - but the reaction of this crowd to your comment is bizarre, and concerning.
Wtf HN. Any of ye silent downvoters want to outline where you’re coming from?
I thought you brought an important inspection to what I had overlooked. And I don't think it's reactionary to find the use of that image offensive. You're right that CS needs to use art with a little more care because we aren't insulated scientists - we are part of society and culture at large.
I think you're being downvoted because it wasn't clear at all what you were talking about, if someone is not familiar with the picture. It just seems like you went on a random rant. Your follow up comment was much more informative about your point of view.
I think it's one of the lowest aspects of our modern civilization, and engineers and scientists simply cannot be ostriches anymore. Though again, I actually think they may have meant it as some kind of "saving the child" thing... I just cannot imagine them working on a high res version of that photo and just making a funny. I don't think that for a second. But in a way, that they maybe thought this was a nice thing to do, makes me sadder. I'd rather be angry and offended, than just... at a loss, you know?
> "I no longer love blue skies. In fact, I now prefer grey skies. The drones do not fly when the skies are grey" [..] Being afraid of a blue sky is the inevitable trauma that comes with the recovery process Zubair and his sister are going through — the two were injured when a drone attack hit their home, and blew their grandmother to bits while she was working in the garden.
At first I thought it was something that could be done in real time. Like, you walk up to any painting with the AR glasses on, do the gesture and it would extract a figure and animate it. On closer inspection, the artworks are also virtual in their example videos. I was hoping they would be real, it feel less impressive when none of it is really present.
This is a series of known techniques assembled in a visually engaging manner, but is ultimately just a dumb gimmick.
* Ability to go from image to model - multiple open source options
* Ability to to from image to rigged figure model - multiple open source options
* motion capture animation playback - trivial
Unless this is instantaneous, which it is not, it's an attention seeking toy with no original ip.
It's amazing but also creepy because characters still retain the non-animated face and/or limbs like in Messi example; making faces moving seems to be logical next step for this technology.
I was surprised at how quickly I went from slack-jawed amazement to wondering why the creators were so enamored with making a character run out of the frame that they never demonstrated using this to animate in context.
I have a feeling that is because they are computer image developers and not animators. I am in the same boat and have a great collection of animations I pull from, but animation is a serious specific skill. Even hooking it to mocap would be another crazy level. Most mocap is gross and needs cleanup, also it would be a distraction from the point. In addition to all of that, by using only a few specific animations it makes the examples easier to compare. Things that look like confusing magic can be hard to quantify on presentation.
35 comments
[ 3.2 ms ] story [ 36.6 ms ] threadThis is a really cool technique that posing interesting questions about how 3D resources in the future will exist.
Will marketing companies release models and textures like we release different image formats now? Are there grest uses for quickly making video or 3D from a photo. IDK, really cool though.
The supplement mentions this about backside texturing:
``` Texturing the back of the body is more difficult, as we have no direct observation of it. One approach is to sim- ply mirror the front texture onto the back. This mirrored- texturing produces reasonable results in some cases (e.g., arms), but undesirable results in others (face appears on the back of the head). To address this problem, we allow the user to choose between mirrored texturing or label-driven texture synthesis – “texture-by-numbers” – on a part-by- part basis. ```
[1] https://grail.cs.washington.edu/projects/wakeup/files/photo-...
Yes it is.
https://en.m.wikipedia.org/wiki/A_Hard_Day%27s_Night_(film)
I haven't read the paper, but am curious how they get the face profile right (given the information is absent in the 2D source). Perhaps they set that manually or some kind of average face, but if it's off it would be quite noticeable.
https://www.youtube.com/watch?v=G63goXc5MyU&t=17s
That's probably meant to be cute or optimistic, I know they're not making fun of anything, but I still find it very tasteless.
Why not just use your own drawings and photos? I generally don't like how art is often used for these kind of things.
I find playing around with that tasteless, to put it mildly and charitably, and I said nothing about it, after seeing it and puking into my mouth a little -- if walking on egg shells, and calling it tasteless to be generous, is punished anway, I may as well be honest, and I'm still biting my tongue on just how tacky CS meeting art is 99% of the time, that's tasteless too but in the actual meaning of the word, just unskilled, garish, time and time again demonstrating that developed art comes from developed persons, not from developed tools -- right after the story was posted and there were no comments. I just shook my head basically, kept it to myself. But when I later saw the question of how long it'll take before someone does something inappropriate with it, I just had to point it out.
What guideline was I violating? How was my comment not a contribution? If you disagree with it being tasteless, if I overlooked or misrepresented something, where are the words?
With a very charitable interpretation it could be said that it’s a warning, a harbinger of the callousnness with which this tech will be used - but the reaction of this crowd to your comment is bizarre, and concerning.
Wtf HN. Any of ye silent downvoters want to outline where you’re coming from?
https://imgur.com/NKHQ2M6
I think it's one of the lowest aspects of our modern civilization, and engineers and scientists simply cannot be ostriches anymore. Though again, I actually think they may have meant it as some kind of "saving the child" thing... I just cannot imagine them working on a high res version of that photo and just making a funny. I don't think that for a second. But in a way, that they maybe thought this was a nice thing to do, makes me sadder. I'd rather be angry and offended, than just... at a loss, you know?
https://www.theatlantic.com/politics/archive/2013/10/saddest...
> "I no longer love blue skies. In fact, I now prefer grey skies. The drones do not fly when the skies are grey" [..] Being afraid of a blue sky is the inevitable trauma that comes with the recovery process Zubair and his sister are going through — the two were injured when a drone attack hit their home, and blew their grandmother to bits while she was working in the garden.
https://notabugsplat.com/
What’s the point of publicly releasing a paper with no real proof of work? They could be faking the algorithm.