I'm searching for Depth map algorithm for action scrip 3 (or some other language).
My goal is to generate depth map for static image on run time.
I have fount this example http://kode80.com/2009/06/29/papervision-zbuffer-and-dof/, i didn't have the time to investigate it's code but i have the filling its working only with 3d models.
Worst viable option is server side technology that does the work.
Extended description:
I have a task to develop flash application where 3d rotation of a head is simulated by a 2d image uploaded by the user.
There are some japan guys that have managed to do that (http://hairtry.jp/), but the technology that they are using is server side. It generates swf fail with precisely coped eyes and image for the depth map of the uploaded picture. I can't find out what is the sever side technology that they are using so I'm searching for any posible workarounds.
Thanks in advance.
Cheers,
Chavdar Slavov
your easiest solution would be to use a 3d engine, Paper vision seems to be getting out dated, theres a lot of documents on how to use, and tutorials but again not much has been happening with them lately. i suggest Away3d "www.away3d.com" they seem up to date.
for what your asking it is possible to do this 100% in flash but to create a 3d head at runtime and fill the face with hieght maps this is possible, you would need to look at bitmap manipulation.i dont think this is an easy task, what you should learn first is how to add an image to a spheric surface dynamically than look into height mapping by converting certain coulours from the image into different levels of depth i must say im sure this is advanced stuff, i manged to animate a globe that was this same method it looks dodgy but it gives you some source code in the FLA
http://www.ffiles.com/flash/animations/globe_animation_3496.html
The above link will allow you to download my source files, the animation it self is a scrolling image but i used Actionscript to give the globe depth and a sheric look
Related
So I have just stumbled upon some augmented reality stuff done in Flash and it got me super excited. I am just looking for some pointers right now. What I would like to be able to do is overlay images on a webcam at specific GPS coordinates. The couple of libraries that I've seen (FLARManager and IN2AR) appear to be based on motion tracking which isn't exactly what I'm looking for. Are there any AS3 libraries that can help do this? Many Thanks.
I'm assuming the key to your project is to get the user's GPS coordinate, and based on this to 'do stuff'.
If you use the native geo-location classes, you can already approximate the users' GPS position: http://help.adobe.com/en_US/FlashPlatform/reference/actionscript/3/flash/sensors/Geolocation.html
However, on a desktop it will not work.
For desktop, you can use a Javascript approach in the HTML to feed the data back into the Flash app: Geolocation in Flash
Then just do your image overlay code on your webcam image as required (though if you don't require motion tracking, I'd be interested to know what sort of 'augmented reality' your project is doing!)
I have for a really long time been searching for a tutorial for an augmented reality application where the users can rotate and\or move a rendered 3D object with hand movements. I have found a few example applications of this, but none with viewable sourcecode, or in any way allowing me to change the models loaded.
What I need is either a working application that I can tamper with in AS3, or a decent tutorial for creating this kind of motion tracking. (I have been able to create a motion tracking application in itself, but not as an AR application where the models loaded are the ones affected by my movement.)
I am rather new to augmented reality programming, and only have the basics of marker based AR down, but I really don't need anything fancy, nor do I particularly need to understand the cone 100% (though obviously it would be better if I did). I just need it to work.
What you need is FLAR toolkit.
http://www.libspark.org/wiki/saqoosha/FLARToolKit/en
Here's a tutorial: http://mlab.taik.fi/mediacode/archives/1939
Enjoy
I was taking a look at this example, http://gardengnomesoftware.com/samples/pano2vr_3/mozart/ and I was really interested in how it was done.
Do you think it would be complicated to develop such a function using one of the webgl engines out there? Is the panorama image just a texture on a cylindrical model? How is the panorama "cut?"
Perhaps it is just a cube with the images as a texture? You can kind of see the faces when it is loading slowly.
Yes, in this case it is simply a cube that the "camera" is centered in. It's basically the same concept as a skybox in many video games.
How to create a cube-map from photos the way they've done isn't something I have experience with, but the wikipedia article links to some tools that may help, including Hugin
Is there a way to embed Sketchup models on a website? Preferably a free one, if it's a commercial solution I'd like a confirmation that it's actually worth the money... I know about the 3D warehouse thingy where you can rotate it (javascript and some images), but I can't upload anything to the 3D warehouse 'cause it always "resets the connection".
You can try the upcoming WebGL Functions in modern Browsers. There are some Javascript Libraries that can read model files and show them within your (HTML 5?) website. I have not used them yet, but try these links if this sounds interesting:
https://github.com/mrdoob/three.js/
https://github.com/timjb/sketchup-threejs/
I made a patch of an example game in three.js. Added ColladaLoader and Sketchup model + Third Person Shooter controls:
https://github.com/irony/Nemesis
Try it out (Works best in Chrome)
http://peaceful-woodland-4410.herokuapp.com/index.html
You can use http://sketchfab.com, they even have a SketchUp exporter.
Yes, you can do it without cost. See here: https://help.sketchup.com/en/article/3000258
EDIT:
There were some free rendering sites apart from Google, but they don't seem to be in operation anymore. Posting to the Google 3D warehouse is going to be your best bet. Regarding your problem posting there, try it with Chrome.
EDIT #2:
If you don't want to upload to Google 3D Warehouse, you can achieve a similar effect with a little work. 3D Warehouse doesn't actually give the visitor the power to manipulate something in 3D, but rather it lets them "rotate" the image horizontally by simply showing a series of preset images of the 3D from various angles. You can take shots of your 3D object, import to Photoshop and create your own gif of the spinning object. A little more work than relying on Google, so you'll have to decide if it's worth it.
EDIT #3: If all else fails you can simply post the 3D model on your site as a downloadable file and advise visitors to get Sketchup to view it.
I am a flash developer who has worked mostly on small ActionScript projects, in particular websites, but I am dipping my toes into image manipulation with AS3. I have a project I'm working on to develop an application which would allow a user to upload their own picture into a swf, and then, using a slider, they can alter the image to make themselves look thinner or fatter. The condition is that it should require minimal user input - so ideally the most I could expect from the user would be a couple of control points identifying certain areas of the face in response to prompts e.g. 'click on the centre of the chin' and when the user clicks, the position would be stored, but this would need to be used sparingly, if at all.
So far, I have been able to use the FileReference class to handle the upload, and I am experimenting with using the DisplacementMap filter to edit certain parts of the uploaded image, but I am having serious trouble implementing this in a manner that would allow realistic deformation of the facial features in the image uploaded.
Is there anything that exists that might run along the same lines and that provides an explanation/source code I could look at to get a better idea of how to go around doing this? Or if there are any other ideas on how to manipulate and distort the image that doesn't use the DisplacementMap Filter?
Any feedback and help is greatly appreciated. Thank you for your time.
see http://sakri.net/technology/flash/flex/convolution_filter/ConvolutionFilterExplorer.html
(source code available)