Camera Emulation

Camera Emulation

Hello, I would like to know if there's an emulator so I can use the SDK until the Creative Camera arrives to my country, because it's really hard to get things imported and it takes a lot of time. If not, I was hoping you could tell how does de SDK get the information from de Camera, does it subscribe to OS events? So I can try to make a Proof of Concept until the camera arrives.

Thank you very much!

8 posts / 0 new
Last post
For more complete information about compiler optimizations, see our Optimization Notice.

What do you plan to do? Someone can save some sequence to a file and you can emulate the camera from the file.


I'm trying to learn how to use the SDK but as I don't have the camera yet I wanted to know if I could test it in some other way. When you say save a sequence to a file is it like a video file? Could you give me an example??

Hi Javier,

>>...When you say save a sequence to a file is it like a video file? Could you give me an example?..

Yes and a webcam or a Point-and-Shot camera could be used to create a video. As soon as the video is created in some format it could be used for simulation. The only problem is that you can spend ( some or significant, etc ) time on that R&D and by the time your simulator is ready the Intel camera could arrive to your country. So, the time could be wasted.

Sergey - what is the best way to answer Javier's question?  The SDK will not work with other cameras.  There is no established format for videos which contain depth maps, confidence maps, and UVMaps.  (We need that BTW.)  Webcam's won't work with the SDK and they don't have depth maps anyway.  My best idea is that Javier should download OpenCV and use it with the Microsoft Kinect camera to learn the basics of using images and depth.

Bob Davies

Bob - that is simply not true, the SDK works with other (depth-)cameras, you just have to provide the data in the same format. The format is described in the common util_* sources. Indeed it will need some work to provide a uvmap too, but for basic gesture recognition you don't need an uvmap.

But was Javier and for sure also many others need is an example sequence that can be downloaded from the SDK website. Only a short sequence where a hand forms a big5 or so and triggers the gesture event would help many developers to do some tests and coding already without hardware!

That is the same point i am now! I was thinking about if we could persist the perceptual inputs and generate some mockups. This ll permit us to explore some sdk features!

Bernardo Rosmaninho 

Bernardo Rosmaninho

Javier G,

you can still use the SDK without the Intel camera if your Proof of Concept is about face detection, landmarks and recognition and attributes only. Hand gestures require the depth map. Here is a sample code with some features that I mentioned ( that should work on a normal webcam.

Leave a Comment

Please sign in to add a comment. Not a member? Join today