Two ways I've used before (and I'm not saying they are the best but they suited my needs) was to use a a jit.lcd to draw a mask onto an image that would be shown on a jit.gl.videoplane - this allows both image and mask to exist on the same plane so they can be moved about in 3d space. Alternatively if you want the image and mask to be seperate (i.e. the mask on a plane in front of the image so both can move in 3d space independently) you could draw a mask into a jit.gl.sketch. Look at the help for both of these and they should get you started.
In view of keeping the forum tidy, here's what I posted on your other thread which is what I think Garrett is talking about in that you can replace jit.gl.gridshape with jit.gl.sketch and draw your own mask instead of using a shape. GL stuff can be tricky to get your head round for a novice so check out the tutorials if you are having trouble.
If you want to use a video of a person as a mask then here's one way to do that using the 'binary-man.mov' footage attached (stolen from this forum as it makes a good test video for just this sort of thing!). To get this sort of footage you can either shoot people against a blank background (or greenscreen) and do some processing which can be tricky to get right without good conditions, or do something a bit more clever with the Kinect (Zachary Sidess' tool can help you here: https://cycling74.com/forums/sharing-kinect-depth-sensitive-blob-tracker-3d-bg-subtraction-etc)