Field3D

Sony Pictures Image works is working on an open source voxel storage format, Field3D.  You can check out the project or the programmer’s guide.

GPU Illumination

Went to the course titled Advanced Illumination Techniques for GPU Volume Raycasting, fortunately it wasn’t a rehash of the 2006 book, but had mostly new work, including some really nice ambient occlusion, scattering, and shadowing techniques.  Also showed some of Voreen.  Really nice bunch of guys.

Ok, I’m off to see a panel discussion with Jenny Sabin of LabStudio.

Gel Sight

Gel Sight is a retrographic surface imaging technique that was wonderfully elegant in it’s simplicity and effectiveness.  They also gave out free samples…

Cuda raycasting 13GB of cryomacrotome goodness (in stereo)

Nvidia had a stereographic interactive realtime rendering of the full 13GB Visible Human dataset being rendered in CUDA on 3 Quadroplexi.  Very impressive.  The glasses used were the new Nvidia active shutter glasses, and were very effective.

Resistive multitouch in many form factors

A new startup out of NYU showed a novel resistive multitouch device.  Very effective, low cost, and suitable to many applications.

UPDATE: Sorry about the broken link, Touchco was bought up by Amazon, so pretty much all of the cool applications they had in mind are replaced by the Kindle 3.

VLC madness courtesy 2 Fusion-io cards

Fusion-io showed their new “budget” nonvolatile storage adapter, the ioXtreme.   $900 gets you 80GB, with a read speed o 700MB/s.  The IO’s aren’t very high, much less their enterprise solutions, but that doesn’t matter if you are reading sequential data.  The booth was pretty crazy, too, one of the better live hardware demos I’ve seen in a while.  I’ll get some pictures tomorrow.  VLC never looked so impressive…

3D Vectorscope Fun!

Stuart at Eyeon gave me the inspiration to try a particle approach this morning after a comment he made regarding my wish for more resolution for the 3D Histogram SubV in Fusion.  Basically, he asked if the size of the cubes was important.  Did I need to know the clustering or not?   Would a 3D Vectorscope be useful?

Oftentimes, just knowing what is or isn’t in the image is enough, and I don’t need to know the distribution.

Took about 20 minutes to get it working, but took a bit longer to make it all pretty and fast and blog worthy.    Bit like the earlier centroid post.

On top is the result, below is the source.   The camera is tracking the result automatically.  It should be possible to use the min and max values to also set the zoom of the camera as well as it’s position to enclose the bounding region and set the camera perpendicular to the longest axis.

3d_particle_vectorscope_c_a07_.comp

(I’ve updated the comp to A07, which adds in a SubV!)

With 256^2 samples, it runs at ~10 Hz on my computer.  1280^2 samples runs at ~0.5 Hz.   Not terrible, but proxy helps a lot.

Couple neat things…  You can clearly visualize the effect of color depth on an image.  8int vs float is quite dramatic.   And you can see how some tools operate very plainly.   Look at a BrightnessContrast for example, and you can see why it’s really an AddScale.   The sample “footage” in the comp is a merge between a Plasma and a BG gradient.  The Plasma makes a ring, while the BG makes a saddle, but together, they make some crazy swirling thing.

You can render the results, too.

Trying out some new datasets and new techniques…

EDIT:  Jim asked for some more details, and I already had some images that I intended to post, but forgot about.  So here’s a breakdown of the three layers used to make the above image…

Chameleon rendered layers

The left layer is an environment map lookup, the middle is a front lit with high opacity, and the right is a backlit with low opacity.   These were then additively composited together.

I also did some tests on this dataset with clipping.

Chameleon culled with spherical gizmo

Chameleon culled with box

The box culling was an accident, but I thought it looked like a cut of meat that had been chewed on by mice.

Here are some real x-rays to use for comparison while we look some commercial artwork.

Example 1: Stylized

   The classic glowing green xray look is show here in this classic clip from Total Recall (1990). 

A close up…

   This was quite a nice effect for eighteen years ago.  It looks quite cartoonish today. Yet almost nothing has changed, except the greenish tint is replaced by a glowing blue. 

This example is even less effective as an x-ray. First of all there is a light source as if this is a regular photograph.  Then there is an unrealistic shader applied to the models.  The opacity of the bones is controlled not by the density of the medium, but by the normal of the surface perpendicular to the camera.  Also there are no other tissues modelled here, only bones.

Here we have another blue illustration. The models are even less accurate. And the shading is also unconvincing. Yet if you browse through medical illustration catalogs, this style is widely used. It’s cheap and quick for the artist to apply and let the viewer see inside.  Yet I doubt that an airbrush artist would make the same decisions in the jaw and eye sockets as these x-ray shaders.

Example 2: Retouched Film

Creating more realistic x-ray renderings in typical software packages isn’t easy.  So clever advertising campaigns might use real x-ray films and digitally pose the person into an activity. Or you can convince people to subject themselves to radiation in the pose itself, like Wim Delvoye.  While the individual bones look much better than the previous example, the artist is transforming the bones in only two dimensions. It’s also quite easy to create incorrect poses for human joints. 

So that’s a taste of the type of problem we work on.

In a later post I will go into the details of what happens in a real X-ray exposure and how to recreate this phenomenon in software.

Thanks,

Ben Lipman
Anatomical Travelogue R&D