4 ms·
Your point about OS interoperability makes sense but you should really, really rethink the "to me kinect is just a sensor, and all I want from it is data" part.
by bignaj 12y ago
Your point about OS interoperability makes sense but you should really, really rethink the "to me kinect is just a sensor, and all I want from it is data" part. If you've ever used a Kinect or similar device half the glory of it is the software processing, voice recognition and face recognition etc. Sure you could probably get just the data if you wanted, but dumbing down the device is certainly not the future. You could have had your version of "the future" ~10 years ago -- just go buy a cheap camcorder. Perhaps you're saying you'd like to write your own audio/video/recognition/processing algorithms for the Kinect?
- colincsl 12y agoAs a robotics/computer vision researcher I disagree with you. The first Kinect was a godsend for robotics - previously it cost thousands of dollars for similar types of sensors. Typically we just want the raw data so that we can do whatever processing we want. It was nice because it worked on Mac/Linux/Windows without too much hassle. The new Kinect on the other hand is of little use to me right now. It's locked down on Windows - and as far as I know you need the special dev kit version for it to even work. Also, a "cheap camcorder" is nothing compared to this. The Kinect is a 3D sensor. You would need 2 "cheap camcorders," stereo vision algorithms/hardware to process it in real-time, and even then the quality typically wouldn't be as good as the Kinect. Tangentially, it's also disheartening (at least for the time being) that Apple recently bought Primesense - the maker of the original Kinect. Primesense was selling a better 3D sensor for developers/researchers. Apple shut it down as soon as they bought them.
- Qworg 12y agoYou're in luck - we're working on it. We've had alpha level success in extracting depth data thus far. https://github.com/OpenKinect/libfreenect2 https://github.com/OpenKinect/libfreenect2 We're always looking for more contributors!
- sp332 12y agoSure, but what happens when you get bought by Apple? :)
- rasz_pl 12y agodoes alpha mean 'unknown binary blob' or 'raw depth data of more or less known structure'?
- ivoflipse 12y agoHopefully this will also mean decent Python support, because with the Kinect for Windows v2 dev kit I'd have to wait until wrappers became available. The Kinect v1 was also a pain to get working on x64 Windows, so hopefully things will be a bit easier this time around
- Qworg 12y agoWe like Python - we support it in libfreenect, so we'll probably have it in libfreenect2.
- thenomad 12y agoHave you tried the Asus depth camera? I work with motion capture apps and 3D scanning, and in both areas developers are pushing it as a decent if not spectacular replacement for the Primesense camera.
- blktiger 12y agoThe connect is a lot more than just a cheap camcorder. It has a regular camera plus a camera for depth sensing and a microphone array (i.e. more than stereo audio).
- sjtrny 12y agoThe software is far from perfect. From a research perspective we just want the data. The reason why the software works so well is because it is a lot more than a colour video camera. It captures depth images too.
- nailer 12y agoI think you might agree with your parent more than you think. Capturing depth is a result of the software - combining IR with stenography. 'Just give me the (high level) data on any OS' seems reasonable. With, of course, access to lower level stuff if you want to create your own depth algo for example.
- darksim905 12y agoAt the end of the day, though, all of that "software processing, voice recognition & face recognition" is just data, as iandanforth said. The Kinect just does stuff with said data :)