Extract Features from Image with Kinect for Windows

I am trying to extract the 3D coordinates of an object using the Kinect for Windows camera. The object to be extracted will be highlighted either with a vivid color, or graphic that can be recognized using feature extraction.

I have used MATLAB and Kinect for both feature extraction and 3D co-ordinate extraction before, however I am having a problem trying to merge the techniques together.

A really simple way to highlight the object, and extract it as a feature in a 2D image is to draw 5 circles really close to each other on the object, Then put it view of the camera and use

CenterCoords = imfindcircles(image, [15 20]);

to extract the co-ordinates of all circles (within a radius range) in the image. You can then remove all circles in the image that don't have 5 similar circles in there immediate proximity, and extract the co-ordinate of the center circle.

But how can I extend this idea to 3D? How can I extract the 3D co-ordinates of a highlighted object using the Kinect Camera?