Gesture recognition and animation

Hi all,

I am planning to make an interactive installation / video wall as a graduation project but I’m complete noob in programming. My idea is to make e.g. 4 different animations with After Effects and use them in an interactive installation that allows people to influence which animation plays using Kinect and gesture recognition. I’m wondering if I should buy Kinect 1 or 2? I know the basics of Processing but should I look into Processing or OpenFrameworks with this project? Also I would be extremely grateful for all kinds of tips on where to start! Thanks!


This is a difficult question. I only just write my(!) opinion on this shortly, hoping it will help.
I myself switched from processing to openFrameworks because of speed mainly. Processing is slower than openFrameworks (although I heard that processing 3 has made quite a step in performance).
On the other hand, as much as I love openFrameworks, I think processing is much easier to understand for beginners. You open it, type and compile. So my a advise - for a “complete noob in programming” as you describe it - would be to get a computer with enough power and use processing.
Getting the kinect to run with processing or openFrameworks is quite easy. You should find a lot of examples and help. I did not use a kinect 2 yet, but be aware that you will need USB 3.0 for it. Also I’d strongly advise to search the web for examples and problems with kinect 2. Kinect 1 has been around for a long time, so you will find more help and examples, probably it’s the better choice, if you are a beginner. With a kinect 1 you should in any case get a model number 1414, not 1473. The latter is very difficult to get to work.

Good luck with your project!

P.S.: If you want to get a bit deeper into the topic of kinect I can advise the book “making things see” by Greg Borenstein.

Thank you so much for your help @dasoe! I ended up buying Kinect 2 (Xbox One) as it was better available in my country. However, now I’m wondering if I should install it using Windows or OS X ( I have a bootcamp in my Macbook Pro), which one would use recommend? Thanks for the book tip, it seems very useful indeed! Do you happen to know if I can use the code examples in the book also with Kinect 2?

You must test.

That’s the easier way to find out which one is the most useful for your projects.

Hi Sofia,

Although cgiles is right that in the end you must (and will) test, I am not very happy about posts that do not add any insight on your topic but lecture nevertheless. I can see nothing wrong in asking for advise and you are specific enough. I do the same here in the forum. Because there are people who have gathered experience already and I maybe just started on a specific topic. Sharing is how it works, so don’t be discouraged. (And maybe, if you start being more experienced, you remember and start sharing the wealth (your knowledge) also. Don’t be afraid to do this also, OK?)

I would recommend using the Mac. I am working with all 3 Systems: Mac, Linux and Windows. I also use the kinect on all three. On Mac I tried first, I found explanations and it was just easy. Later, after gathering some experience I tried Ubuntu. It was still easy, but Installation (and understanding what’s happening) a bit more complex. Last I did it on windows and eventually succeeded. But although having much more experience by then, it was not as smoothly as on the other systems.
On Mac on the other hand you should use xCode while on the other systems (at least on linux) you are more free. But again, as a beginner, I think there is no problem on not having too many possibilities.

I would not bother about the code in the book being exact. If (just in case) something really has changed (over time or because of the new kinect) you will find out in the web. This will not hinder you to get to know the kinect and to understand what’s happening - which is what the book is meant for.

have a good time and build a great project!

To use windows or mac for kinectV2 depends on what features you want. As yet there is limited support in mac for this kinect, there is no skeleton tracking and there is a robust but primitive version of an addon that can get the raw streams of IR depth and colour- it does however allow for multiuple cameras.

On windows you can access the proper SDK through openframeworks and it is quite advanced and stable, giving full features like skeleton tracking etc.

But in summary you must test what you need.