Nsquared "Seamless Computing demo" shows the awesome things you can do with MS Surface, a tablet, ki - Polidicks

[QUOTE=TheCloak;32024458]Seems like a tool that criminals can use to plot escapes without having to actually scout the place out first[/QUOTE] You don't need sci-fi animations, hand gestures, and a voice activated 3d model viewer to look at a floorplan. [QUOTE=icemaz;32025285]I like how you're al focusing on the bit where he makes a 3d model from a picture which seems relatively reasonable considering he just scanned a piece of paper with a screen.[/QUOTE] Scanning a piece of paper is entirely possible. Generating a detailed 3d model of a lamp using one 2d reference image is not.

[QUOTE=Wnd;32026051]Regular multi-touch screen can detect shape of objects easily. It is more of software.[/QUOTE] Yes but not with the accuracy of Surface. The pixels have IR sensors that can detect shapes and colors. It's all hardware.

They stored blueprints of every house in the world?

Lots of eye candy. Not that that's a particularly bad thing...

If we could make 3d models from pictures there wouldn't be a point in modelers anymore unless the object didn't exist.

[QUOTE=Doctor Zedacon;32026149]Apparently some people from the Unlimited Detail following wandered in here and got offended because I made a very correct comparison between the two topics. Well done. It is actually really humorous.[/QUOTE] While your comparison is very misinformed, I will have you know I think both this technology and UD are very possible, and support both. Considering the 3D scanning from a 2D image thing probably just contacts a database. If it's not though, it's completely impossible unless the program makes assumptions, because there isn't anywhere near enough depth information stored in a 2D image.

[QUOTE=NorthernGate;32026492]While your comparison is very misinformed, I will have you know I think both this technology and UD are very possible, and support both. Considering the 3D scanning from a 2D image thing probably just contacts a database. If it's not though, it's completely impossible unless the program makes assumptions, because there isn't anywhere near enough depth information stored in a 2D image.[/QUOTE]The assumption though is that it only took a 2D image, which is not stated at all in the video. It is entirely possible that it took a 3D image and use the data collected from both cameras to help create the model. And I'd really like someone to demonstrate how my comparison was misinformed.

Is it an Australian custom to demo incredible technology and not show how it works? (first UD and now this).

Manchester has a Microsoft Surface, I'm going Manchester tomorrow, might have a look. The one in Manchester is only used for Visitor Information, still cool though

Wow that was really redundant and unnecessarily complicated.

This is just a preset show off video. I'm sure it's nowhere near as useful as it looks.

I was laughing my ass off at the "rendering" bits

[QUOTE=Doctor Zedacon;32026594]The assumption though is that it only took a 2D image, which is not stated at all in the video. It is entirely possible that it took a 3D image and use the data collected from both cameras to help create the model. And I'd really like someone to demonstrate how my comparison was misinformed.[/QUOTE] It's misinformed because UD isn't impossible, its been done many times before. If you want papers on how it works and some videos of other companies doing it look up "sparse voxel octrees" because it's the exact same thing. So thats true, thats fact, and can be demonstrated as such. On to this video. Theres no back side of the lamp that's ever seen, only the 1 angle. Even if it could grab depth information and uses data from both cameras, it wouldnt have enough information (or resolution) to build that model of the lamp. It's impossible. To build the data behind, the computer would need to make judgements about the missing data. And a database of objects wouldn't work either since there's just far too many objects in the world to even bother making models for all of them, and searching a database by data from an image is a difficult task. A little off topic but for anyone interested, some videos of other sparse virtual octrees, to show that 1) Its not impossible at all and its a well documented technique, and 2) the guy that presents the videos is a smug bastard, and this isn't some sort of secret company algorithm he invented in his laboratory or something: Made by a guy at ID to see if if it's feasable for IDs next engine or so: [media]http://www.youtube.com/watch?v=VpEpAFGplnI[/media] And one that NVidia put out using cuda: [media]http://www.youtube.com/watch?v=lpfaFrazOn4[/media]

Stop talking about Unlimited Detail plzthx.

[QUOTE=Doctor Zedacon;32026594]The assumption though is that it only took a 2D image, which is not stated at all in the video. It is entirely possible that it took a 3D image and use the data collected from both cameras to help create the model. And I'd really like someone to demonstrate how my comparison was misinformed.[/QUOTE] It is not possible. UD isn't all it's chalked up to be. Stop commenting on stuff you don't know about. show me proof or ~buzz off~

i was more interested in the cool phone-on-table screen thing than the bullshit house layout and object rendering tbh the way he just threw that paper on the table and scanned it was cool as hell

[QUOTE=Doctor Zedacon;32026594]The assumption though is that it only took a 2D image, which is not stated at all in the video. It is entirely possible that it took a 3D image and use the data collected from both cameras to help create the model. And I'd really like someone to demonstrate how my comparison was misinformed.[/QUOTE] There was only one camera on the tablet. You can't make a plausible 3D image with one camera, on a handheld computer no less.

[QUOTE=Kinglah Crab;32027312]i was more interested in the cool phone-on-table screen thing than the bullshit house layout and object rendering tbh the way he just threw that paper on the table and scanned it was cool as hell[/QUOTE] Windows Phone has that kind of integration with Surface natively. God I love WP7.

You know, i think this is all a sham. The lamp that other people have mentioned - magic accurate 3d information from a 2d image, plus how did the computer know what the lamp was when he said "put lamp on shelf". It even knew that the inside of the lamp was hollow which [i]it couldnt possibly have known under any circumstances[/i]. And the shelf for that matter, how did it know without prior knowledge of it that it was "shelf"? The house looks like it has been programmed to interact with him like that, i severely doubt the technology works like this in practice without some [i]serious[/i] amount of information being given to the computer beforehand, like voice samples for ever command you give it, dimensions and placements of all of the objects in the house etc. The piece of paper document at the end, it must have digital information on for the computer to read. There is no way im going to believe a computer read the document and identified how to correctly order a lamp and the doorhandle without prior knowledge. Tl;dr, its possible but not without informing the computer of every little step you make before you do it

[QUOTE=Doctor Zedacon;32024845]Kinda funny though, seeing a lot of people jumping to say "No! That's bullshit! You can't do that!" But I'm gonna have to point you to that little thread nearby about Euclideon's Unlimited Detail.[/QUOTE]"Hey guys people said bad stuff about something completely different and it turned out fine so obviously that means that this thing completely unrelated is good because some people are calling bullshit"

[QUOTE=NorthernGate;32025649]im pretty sure that'd break the screen[/QUOTE] Ever heard of hyperbole

its CSI:NY all over again [editline]30th August 2011[/editline] nullsquared would be so happy [editline]30th August 2011[/editline] They are overdramatic with the effects [editline]30th August 2011[/editline] And the lamp thing won't work without 2 cameras

This seems really bloody useless, especially when almost everything he showed in the video would require data detailed to the nth degree, which realistically isn't going to happen in a commercial marketplace

"Bill, show me your penis." "Bill, create a new world completely made of living couches that reside in South Africa." "Bill, MOVE THE FUCKING LAMP!"

i loved the high-tech sci-fi effects they added

My local library has one of these, though it only has a very strange game on it, and barely recognizes your touch. They also use one of these in both hawai Five 'O and NCIS:LA

[QUOTE=mobrockers2;32029730]My local library has one of these, though it only has a very strange game on it, and barely recognizes your touch. They also use one of these in both hawai Five 'O and NCIS:LA[/QUOTE] I've had the same experience. As I said earlier, you have to press extremely hard for it to recognize your touch. It makes it as whole a very clumsy and awkward system to use.

[QUOTE=OpethRockr55;32024687]The 2D picture could represent 3D data, like a model or such that's loaded up when he drags it to the 'big screen'. The picture of the lamp into 3D is horseshit, though. You can't pick up accurate depth information from a single picture, and yet it somehow even knows there's a button on the back of the lamp head not to mention the correct colors and depth information of the entire object. Unless it only deals with a preset amount of data or just guesses at whatever object is closest to the one you want, it's impossible to obtain that much data.[/QUOTE] The only way I can see it working is if it has some kind of image comparison technology and searches a collection of websites with those applications, finding the correct clamp and passing it on to the big screen.

[QUOTE=SomeRandomGuy16;32029749]I've had the same experience. As I said earlier, you have to press extremely hard for it to recognize your touch. It makes it as whole a very clumsy and awkward system to use.[/QUOTE] In all fairness Microsoft is releasing Surface 2.0 which is thinner and more responsive.

Another demo of it's capabilities: [media]http://www.youtube.com/watch?v=KUFkb0d1kbU[/media]