Thursday, September 19, 2013
This Impossible Software Can Make 3D Models From a Single Photograph
gizmodo.com: It was predicted that computer graphics would one day give everyone the ability to create their own blockbuster film. But the software used for modern visual effects is still pretty complicated for the average user. At least until this magical software that can almost instantly turn a still photo into a 3D model is available.
Subscribe to:
Post Comments (Atom)
12 comments:
There is no way this software works as reliably as it looks in the video and if it does this needs to be on our wishlist. Extrapolating and creating a 3d parametric model from a photo is awesome. This would make the most complex drafting tasks so simple. We would be able to create a complete 3d model of a set from a rendering with minimal additions of obscured objects. This is bordering on star trek technologies.
I'm very impressed with the software that this video shows off. I'm not sure if it actually could work as well as they make it look. I was getting really skeptical as I watched the video and started to with that they would tell me the down sides of this software, and then they did exactly that. They deserve a lot of credit for showing off the short comings as well as the advantages. Something that I see a lot in my life is people who are very impressed by 3D models. I think this stems from their complete lack of knowledge about how 3D modeling works. Many times people would love to have a 3D model but don't know where to start, because they see it as to hard of a task. This software has a lot of potential for helping this kind of person do jobs that they otherwise would have walked away from. I don't know that the software is ready for "real" use just yet based on what I saw at the end of the video, but maybe soon it will be ready for the real world.
I saw this video on a different website a few days ago and I wrote it off as a scam. This is definitely the direction that software is heading, and this looks like a really great start. However I feel like either they are showing us shots where they got lucky, or the shapes the program is making are simpler then the shapes they are attempting to replicate. I think the later is evidenced by the first shot where the cube with the set in square on top is replicated as just a cube. Still really cool, but its got a ways to go.
Since I have started my 3d modeling career and let me say it is not the easiest thing I have ever done. BUT WOW!! This software is crazy and I am in love. Of course, does the software REALLY work that well, REALLY?!?! It is like the Iron Man Cad video from last week but it is still very cool. I would love to personally investigate this software and see what it is really like. Until that great day comes I will stay HOPEFUL that I will one day have a holographic modeling program that can talk to me!!
Yeeah....I don't buy it. Yes, the video is interesting. Yes, it makes it look like the program functions well. Yes, it fakes details moderately well. Yes, it makes a model from a picture...but how well does it actually do those things. Not many of those shapes were incredibly complicated, and the ones that were, the model seemed to match the picture...ish, + or - 5% accuracy, maybe. I'd also wanna know how well this program can export these models, and what file types it can process. Can I export an .stl file to AutoCAD...? This video seems like it shows a decent amount of functionality, but I'm still left wanting.
I think this is really neat, but I agree that they probably only showed us the lucky examples. This shows that the software automatically assumes that an object is a sphere and not an ovoid, or a square and not a rectangle. In some of the examples, they could show that an object was a rectangle because of the angle of the object in the photograph. Most of the photographs showed several perspective viewpoints for the objects. If you have a picture of the same object head on, you can't see the depth. This software is obviously limited as to what you can do with it, but it is definitely a stepping stone.
I agree with Luke and Joe and Lindsay, I don't think this is as perfect and wonderful as the video shows. I think it is absolutely amazing, I just think that it's too good to be true. The failed attempts that they showed at the end of the video made me realize that it probably works a lot less often than when it actually does work. I'm glad that technology is headed in this direction though, its going to be really useful once all the kinks are worked out.
This video shows off some really cool new technology. It is, sadly, still in the research phase so it is unlikely that it is ready to work in the way the video shows in a user friendly way yet. The video also shows some failures of the software in it's current state near the end. Still, this video means that the software is being developed and the algorithms behind it are at a solid enough point that they can be used. Now it is up to the researchers to improve reliability and usability so that their tools can actually be put to use in the real world.
So many people are skeptical and I can understand why. But personally, I think this program will ultimately be very successful. It is currently in the R&D phase and it might be a little while before it comes out, but when it does, I get the feeling a lot more people will be doing 3D design and modeling. For example, I can see it work its way into designing/decorating rooms and homes because you can just take a picture and then add it to the model and see how you like it.
Yes, I know I am thinking waaayyy ahead, but I don't think it is as far off as people think...
This program works so well, it doesn't seem real. The basic concept makes sense and seems possible but the programs ability to sense the object's outline, and formulate a 3d model, replicate the background image, and replicate the objects texture on the other side of it just seems like a little to good to me true.
This is sooooo cool!!!!! I see it working completely. The software that the 2014 Microsoft word and PhotoShop uses to select objects based off of there outlines and ridges makes since for the 3d effect. I feel like the video makes it seem simultaneous which theres no way that it would be, and also it would only work with objects that are completely the same on all sides but the idea behind this is great and i believe will be extremely useful in the future for people like me who dont really know how to CAD or Vector, especially in 3D. The other thing that confuses me about this is the perspective. It seems like the perspective of the picture would effect the actual detentions of the modeled object. They briefly touched on this but for a different reason.
wow what a powerful program. I can see this being used to collage in 3D. But not only that but do it in the fraction of the time usually associated with the task of collaging. The possibilities seem to be endless with a soft wear like this. I could imagine taking my own hand drawing renderings and scaling then to fit the venue in scale, or making realistic shapes from the actual sorus. I do wonder a little be if this will have any negative effects on any current corporation or job field. At the very least it looks as though it will help more than hurt, but I do wonder at what coast. As technology gets more and more evolved we start to lose some of the great things as well as make our lives easier. I just hope we don’t have to lose to much to be able to use this what will be a great tool, I’m sure.
Post a Comment