This depends on the type of VR setup you want to use. If you're looking at a setup with a 'fixed' display such as a desktop screen or monitor then the WiiDesktopVR approach is right for you.
If however you are looking at achieving a cheap head tracking solution for a head-mounted display you'll need a different approach since it is not reasonable to assume that the user is facing the tracking wii remote (therefore the IR sources won't always be visible).
One approach is to hang the tracking wii remote above the user and track IR points from above. This can measure translation and yaw with just two IR points within the limited field of view of the wii remote.
Yang-Wai Chow, (3D Spatial Interaction with the Wii Remote for Head-Mounted Display Virtual Reality. World Academy of Science, Engineering and Technology, 2009. 50: p. 377-383) used this approach with an additional wii remote (the accelerometer values were used to augment the limited IR point tracking) to create a pointing device within a HMD VR system (using magnetic head-tracking for the user display). They found that jitter impacted performance so they used a kalman filter (which introduced some lag) and that the most accurate tracking was done within 2 meters of the remote (limiting the FOV).
Alternatively sing additional point tracking should be able to provide you with 6DOF tracking.
Then your problem is creating a 3D scene. I've just finished a project that used Brian Peek's WiimoteLib within a pre-existing "DirectX SDK Managed Direct3D sample framework" project. So at least if you used this sample framework I can assure you WiiDesktopVR-style head-tracking is possible