CODEWITHSUNDEEP
computer-visionkinect
thinkquester
thinkquester

Reputation: 326

How to get skeletal coordinates along with RBG + Depth video and save the data using Kinect?

I am working on a Machine Learning project (Action Recognition) that involves capturing RGB + Depth Video data. As one feature, I also need to extract the skeletal coordinates of the joints. Getting skeletal is easy, but does KinectSDK provide such method to get Skeletal from prerecorded videos?

I was thinking of making an application that would record RGB and Depth Videos and save the coordinates in a file frame by frame. Is that the best method?

Upvotes: 0

Views: 1395

Answers (1)

Sassa
Sassa

Reputation: 3334

No, the Microsoft Kinect SDK does not provide you with the capability to get the skeleton from a prerecorded video. Also, it doesn't provide you with the mapping from skeleton data to depth if the sensor is off.

I would suggest writing the mapped to depth (or color if you prefer) joint coordinates to a file when you record your RGB+Depth data. I haven't done it, but I guess it wouldn't add much overhead.

I am guessing you have seen it, but for completeness this is the link to the Channel 9 tutorial for skeletal tracking which includes those mapping functions.

Upvotes: 1

Related Questions