we can achieve gestures using kinect, I am thinking to get images through kinect and then apply neural network on it, to get more accuracy. is it good approach or i am thinking in wrong way?
Asked
Active
Viewed 380 times
1
-
Have you tried https://msdn.microsoft.com/en-us/library/microsoft.kinect.visualgesturebuilder.aspx ??? – Rafaf Tahsin Sep 04 '16 at 06:39
-
there is no neural network involved in it. i need to apply neural network on the stream of the output – naveed shah Sep 05 '16 at 07:42
1 Answers
1
Using neural network to recognize human gesture is a good idea, and some guys have published their result.
Dong Li
- 520
- 2
- 7
- 18
-
@naveedshah I haven't used kinect, but familiar with neural networks like CNN for image classification. – Dong Li Sep 03 '16 at 03:50
-
-
You cannot run CNN on video directly. If you want to build training data set, I suggest you can extract the image from video 10 FPS with the corresponding label(maybe the gesture class). If you try to test your CNN on a video, you must extract a single frame and run CNN on it, then the next frame. – Dong Li Sep 06 '16 at 02:38