Have Geek Brain

Built-in computer vision platforms

I am planning to start a computer vision based project on a smartphone platform.

I know iPhone and Andriod have openCV support. I am interested to know how your experience is with the level of integration, support and ease of building good applications on both platforms.

I also want to consider Windows 7 (and Zune) as a platform. Are there Computer Vision libraries for this platform or any good development tools (does Aforgenet work or any other good suggestion)?

Also you can suggest some popular augmented reality apps that use the latest technologies (I know Pranav Mistry SixthSense )

Thnx in advance!

+2


a source to share


3 answers


Unfortunately, there is currently no access to the camera data stream on Windows Phone 7. Instead, you're limited to launching a camera app, allowing the user to take a photo and then retrieve that image's data.



This is a requested feature, so it could be changed at any time.

+3


a source


Look WordLens for cutting edge AR.



+2


a source


I've created AR and other iOS computer viewing apps using OpenCV and found it to be a solid platform. Otherwise, it provides you with a fast, reliable set of libraries for Matrix math, as well as optimized versions of some of the more common vision algorithms for extracting objects into 3D reconstruction. This is pretty much a defacto standard, so there is a great support community there and I would definitely recommend it.

From a development perspective, I tend to write OpenCV command line apps on my Mac, and then when debugged and working, I look at moving them to iOS. This shortens the test / debug cycle (since I don't have to worry about deploying and debugging on iOS) and also allows me to focus on the problem rather than the mobile quotes.

It also makes the transition to Android platforms easier, although I found the Java wrappers for OpenCV on Android less performant. Again, focus on getting your main algorithm and processing pipeline running on the desktop in (say) C ++ and then moving it to porting the mobile device to the native code format you want.

0


a source






More articles:


All Articles