Getting Started With Kinect.v2 and Cinder, pt. 1: Cinder-Kinect2

...Gentlefolks, start your frameworks.

    So, I'm somewhat ashamed to admit this, but I've actually had my Kinect v2 devkit for about a month-and-a-half now and haven't really done much with it other than run some of the demos, browse the sample code, and in general be amazed at the technology.  Thank god for eyeo festival and time off for a week to just create and be inspired, as it allowed me to really get hands on with the hardware and some of the solutions available for Cinder.  Let's take a look, then, shall we?  In this first installment, we'll go over Cinder-Kinect2, which is a wrapper around the core SDK, developed by Stephen Schieberl of Wieden-Kennedy and of course, Cinder, fame.

---

DISCLAIMER: This is preliminary software and/or hardware and APIs are preliminary and subject to change

---

    Let me preface this by saying I'm not presenting anything show-stopping here, this isn't even Cinder Kinect 101 (it's more like 90), chances are most users probably know this stuff already, but at least it's here for some amount of posterity, and I will be updating this as the APIs mature (which I guess should be any week now?).

    Ok, that all said, here we go, for real.  Cinder-Kinect2 is the simpler of the two solutions, granted, they're both simple, Cinder-Kinect2 just requires less to get started.  How simple?  Well, assuming you have the K4W DPP SDK installed...

-> Clone or fork Ian Smith-Heisters' fork of Cinder-Kinect2, as it includes some fixes to account for changes in the 1404 Developer Preview SDK.

-> "Install" it as you would any ordinary Cinder block, i.e. clone your chosen repo into your <cinder_root>\blocks folder.

-> Pop open a sample, make whatever changes you need, if any, to the Project Properties.

-> Build and Run. ??? Profit!

    As of this writing, the main fork of Cinder-Kinect2 doesn't take the 1404 changes into account.  If you're interested in fixing it yourself, really you just need to make some minor changes involving the commenting out or removal to references to KinectStatus, as it's been deprecated, replacement pending, but the good Mr. Heisters has also added a few other features that make his branch worth taking a look at.

    The interface itself is as simple, if not simpler, than the installation process.  Here's a sample image and a snippet of the code that produced it, cribbed from the BasicApp and BodyApp sample:

void K2TutorialApp::update()
{
  if(mKinect->getFrame().getTimeStamp()>mFrame.getTimeStamp())
  {
    mFrame = mKinect->getFrame();
    if(mFrame.getColor())
    {
      mRgbTexture = gl::Texture(mFrame.getColor());
      auto cBodies = mFrame.getBodies();
      if(cBodies.size()>0)
        mJoints = cBodies.at(0).getJointMap();
    }
  }
}

void K2TutorialApp::draw()
{
  gl::clear(Color(0, 0, 0));
  if(mFrame.getColor())
  {
    mRgbTex = gl::Texture::create(mFrame.getColor());
    gl::draw(mRgbTex,
          Rectf(Vec2f::zero(), getWindowSize()));

    auto cJoints = mBody.getJointMap();
    if(cJoints.size()>0)
    {
      gl::pushMatrices();
      gl::scale(Vec2f(getWindowSize()) / Vec2f(mRgbTex->getSize()));
      for(auto cCjoint : cJoints)
      {
        Vec2f cp = Kinect2::mapBodyCoordToColor(
                cCjoint.second.getPosition(),
                mKinect->getCoordinateMapper());
        gl::drawSolidCircle(cp,10);
      }
      gl::popMatrices();
    }
  }
}

Resources

• Cinder Home - http://libcinder.org
• Cinder on github - https://github.com/cinder/Cinder
• Ian Smith-Heister's Cinder-Kinect2 fork - https://github.com/heisters/Cinder-Kinect2
• Stephen Schieberl's Cinder-Kinect2 fork - https://github.com/wieden-kennedy/Cinder-Kinect2
• K2TutorialApp Source - https://github.com/SethGibson/K2Tutorial

A video is worth 30,000 words per second?

Point Clouds and FBOs from Seth Gibson on Vimeo.

    Just popping in to say that if you haven't checked out my Vimeo channel recently...err...there's not much new stuff there, but there's more fun stuff coming (along with new blog posts that I've promised people and myself).  I'm considering doing some Cinder video tutorials if I ever find some free time, not sure about what, maybe covering Kinect, the Intel depth cameras, that sort of thing...not sure really, we'll see.  But anyway, yeah, bookmark my channel, show your friends, loved ones, your inner circle, all that.  If nothing else, it's good for a tiny bit of inspiration...maybe.

>>> Me on Vimeo <<<

Cinder-Assimp, VS2012, and Cinder 0.8.5

If you're in a hurry, absolutely need to get Cinder-Assimp functional, and don't have time to read my drivel...TOO BAD!  I kid, I kid, here's all you need to know:

• Download the pre-built Assimp 3.0 libs from Eric Renaud-Houde's Cinder-Skinning block and stick them in Cinder-Assimp's lib/msw folder
• Pop open AssimpLoader.cpp and change line 460-ish from cam.setFovHorizontal to cam.setFov
• Shake your head sadly at the fact that I'm so desperate for attention that I took a two step process and turned it into the yarn below, which you may optionally read.  Yep, I'm definitely management material.(Optional)

    More often than not, when we need to deal with 3d content here in Lab land, we tend to use Unity as our framework of choice, though with dropping of the thermonuclear bomb that is UE4 on reality this week, that may be changing.  The problem, as you can imagine, is that sometimes we don't really need something as extensive as Unity, but we end up using it anyway because we don't have options, which leads to some really interesting design decisions sometimes.  Recently, we had a project that really only required a minimal amount of 3d content management, so little so that it would've been nice to be able to keep the whole project in Cinder, but since our content was skinned and animated, Cinder's built in OBJLoader obviously wouldn't suffice.  So we were stuck in a situation where really we needed the absolute most minimal subset of Unity's functionality, but because it was our only option, it required quite a bit of one-off development to turn Unity into something like a useable component.  Well, gee Seth, why didn't you just use Cinder-Assimp?  Great question...

...check out the big brain on the Cinder forums.

    Cinder 0.8.5 is an interesting beast, there seem to be quite a few interface changes from previous versions, that, if the dev docs are to be believed, may be re-appearing in 0.8.6.  This was probably the thing that kept us from designing with Cinder-Assimp in mind originally, the fact that we couldn't really get it to work.  Thankfully, we had some time to revisit the project and wrap it up, so I made it a priority to figure out Cinder-Assimp.  If you're a Cinder user, you're probably familiar with Gabor Papp's work, if not, well, it's good, and if he says it worked, then at some point it did, so I figured it was more an error on my part than the block's part.  Turns out I was right.

    First, all credit where it's due, it was actually Eric Renaud-Houde's excellent Cinder-Skinning block that solved the major problem, that of getting compatible libraries.  Sure, I could've just built Assimp myself against VS 2012, which also would have solved that problem, but hey, I'm lazy and am more than happy to let other people do work for me, although I'll probably be running a build of it myself going forward just to make sure I have it.  As much as I'm excited about UE4, I'm definitely not abandoning Cinder...ever.  I like to think of UE4 (or whatever game engine I use) as the infantry and Cinder as Spetsnaz.    Anyway, if you grab the pre-built Assimp libraries linked on Eric's repo for Cinder-Skinning, you can link the Cinder-Assimp samples against that.  But wait, there's probably one more error you might run up against...

...i know, i know, it's always one more thing.

    This..."fix", as I alluded to earlier, gives me a bit of pause, but hey, if you need to get work done, you need to get work done.  But first, consider this:

• Camera.h master
• Camera.h dev

    As you can see, the missing setFovHorizontal() will be (is) returning (here-ish), so my actual advice is to man up and work out of the dev branch and build your own version of Assimp, which you'll probably want to do anyway since Cinder is also moving to boost 155, so why not just build everything against the same boost?  But again, if that's not an option, you can just patch AssimpLoader.cpp to use the setFov() method vs setFovHorizontal() for now.  So to recap (stop me if you've heard this one before):

• Download the pre-built Assimp 3.0 libs from Eric Renaud-Houde's Cinder-Skinning block and stick them in Cinder-Assimp's lib/msw folder
• Pop open AssimpLoader.cpp and change line 460-ish from cam.setFovHorizontal to cam.setFov

    Now to finish up some prototypes so I can go to a Systema seminar this weekend and not sleep at the office.  Hopefully I'll carve out some time to play with UE4 as well, although, I had an interesting conversation with Stephen Schieberl, also of Cinder fame, at eyeo festival last year where he alluded to the idea that Team Cinder's been looking at game engines and thinking about how to bring some of those ideas, especially on the content creation side, into Cinder land, so...maybe UE4 is just a passing fancy too.  The future is exciting, my friends, go make something.

...in the future, all UIs will be MADE OF CASCADE PARTICLES!!!

The Human Triangulation Experiment, Stage 1

    Recently, us lucky saps in the Perceptual Computing Lab have been fortunate enough to be doing some prototypes for different external groups, and I've actually been lucky enough to work with one of my favorite groups, [REDACTED]!  Needless to say, I'm super excited, and one of the first projects I'm working on involves using depth data and object/user segmentation data to interact with virtual/digital content, much like the ever popular kinect/box2d experiments you've probably seen floating around...

Check out more of Stephen's stuff on github or on his website.

    Depth buffer, OpenCV, cinder::Triangulator, and Box2D, seemed pretty straightforward, I mean let's be honest, that's creative coding 101, right?  That's what I thought, but as usual, the devil's in the details, and after some (not terribly extensive) searching and a little bit more coding I had...eh, well...nothing.  My code looked correct, but no meshes were drawn that day, and even in a cursory inspection of my TriMesh2ds, there was nary a vertex to be seen.  Here's what I tried originally (this is sketch code, so yeah, the pattern is a little sloppy):

//cv contours are in mContours, mMeshes is vector<TriMesh2d>
for(auto vit=mContours.begin();vit!=mCountours.end();++vit)
{
  Shape2d cShape;
  vector<cv::Point> contour = *vit;
  auto pit=contour.begin();
  cShape.moveTo(pit->x,pit->y); ++pit;
  for(/* nothing to see here */;pit!=contour.end();++pit)
  {
    cShape.lineTo(pit->x, pit->y);
    cShape.moveTo(pit->x, pit->y);
  }
  cShape.close;
  Triangulator tris(cShape);
  mMeshes.push_back(tris.calcMesh());
}

    Right, so at this point, it should be a simple exercise in gl::draw()ing the contents of mMeshes, yeah?  Sadly, this method yields no trimesh for you!, and as I mentioned above, even a quick call to getNumVertices() revealed that there were, in fact, no vertices for you!, either.  The docs on Triangulator lead me to believe that you can just call the constructor with a Shape2d and you should be good to go, and a quick test reveals that constructing a Triangulator with other objects does in fact yield all the verts you could ever want, so methinks maybe it's an issue with the Shape2d implementation, or perhaps I'm building my Shape2d wrong.  I rule the latter out, though (well, not decisively), since Triangulator has the concept of invalid inputs, e.g. if you don't close() your Shape2d, the constructor throws, so...what to do, what to do?  To the SampleCave!

TRIANGULATE AGAIN, ONE YEAR! NEXT!

    Mike Bostock, he of d3.js fame gave a great talk at eyeo festival last year on the importance of good examples (Watch it on Vimeo), and you know, it's so true.  It's sorta like documentation, we employ technical writers for that sorta thing, I feel like we should at least give some folks a solid contract to put together good sample code for whatever we're foisting onto the world, rather than relegating samples to free time and interns (no offense to either free time or interns).  Now Cinder has amazing sample code, so a quick google search for TriMesh2d popped up the PolygonBoolean sample, which was basically doing what I wanted, i.e. constructing and drawing a TriMesh2d from a Shape2d...kinda.  I trust the good folks at Team Cinder to not ship sample code that doesn't work, so a quick build 'n' run later and I had a solution.  I was sooooo close...

//cv contours are in mContours, mMeshes is vector<TriMesh2d>
for(auto vit=mContours.begin();vit!=mCountours.end();++vit)
{
  PolyLine2f cShape;
  vector<cv::Point> contour = *vit;
  for(auto pit=contour.begin();pit!=contour.end();++pit)
  {
    cShape.push_back(fromOcv(*pit));
  }
  Triangulator tris(cShape);
  mMeshes.push_back(tris.calcMesh());
}

    The results?  Well, see for yourself:

My tribute to Harold Ramis, may you never end up in one of your own traps, sir.

    Next steps are to maybe run some reduction/smoothing on the contours, although I suppose it doesn't matter terribly for this prototype, and get it into Box2D, all of which I'll cover in Stage 2, including a quick 'n' dirty Cinder-based Box2D debug draw class.  This is awesome, it's total Tron stuff, bringing the real into the digital and all that sort of sorcery.  Once I get the Box2D stuff implemented, I'll stick a project up on github, until then, if you have specific questions, Tag The Inbox or leave a comment below, you are always welcome to try my Shogun Style...for reference, here's the complete update() and draw():

//Using Cinder-OpenCV and Intel Perceptual Computing SDK 2013
void segcvtestApp::update()
{
  mContours.clear();
  mMeshes.clear();
  if(mPXC.AcquireFrame(true))
  {
    PXCImage *rgbImg = mPXC.QueryImage(PXCImage::IMAGE_TYPE_COLOR);
    PXCImage *segImg = mPXC.QuerySegmentationImage();
    PXCImage::ImageData rgbData, segData;
    if(rgbImg->AcquireAccess(PXCImage::ACCESS_READ, &rgbData)>=PXC_STATUS_NO_ERROR)
    {
      mRGB=gl::Texture(rgbData.planes[0],GL_BGR,640,480);
      rgbImg->ReleaseAccess(&rgbData);
    }
    if(segImg->AcquireAccess(PXCImage::ACCESS_READ, &segData)>=PXC_STATUS_NO_ERROR)
    {
      mSeg=gl::Texture(segData.planes[0],GL_LUMINANCE,320,240);
      segImg->ReleaseAccess(&segData);
    }

    mSrcSurf = Surface(mSeg);
    ip::resize(mSrcSurf, &mDstSurf);
    mPXC.ReleaseFrame();
  }

  cv::Mat surfMat(toOcv(mDstSurf.getChannelRed()));
  cv::findContours(surfMat, mContours, CV_RETR_LIST, CV_CHAIN_APPROX_SIMPLE);

  for(auto vit=mContours.begin();vit!=mContours.end();++vit)
  {
    PolyLine2f cLine;
    vector<cv::Point> contour = *vit;

    for(auto pit=contour.begin();pit!=contour.end();++pit)
    {
      cLine.push_back(fromOcv(*pit));
    }

    Triangulator tris(cLine);
    mMeshes.push_back(tris.calcMesh());
  }
}

void segcvtestApp::draw()
{
  // draw camera feed
  gl::clear(Color( 0, 0, 0 ) );
  gl::color(Color::white());
  gl::draw(mRGB, Vec2f::zero());

  //draw meshes
  gl::enableWireframe();
  gl::color(Color(0,1,0));
  for(auto mit=mMeshes.begin();mit!=mMeshes.end();++mit)
  {
    gl::draw(*mit);
  }
  gl::disableWireframe();
}