There are ways to do this, You would a Multi tracked NLE Video Editing Program. Yes it does require masking. What you do is shoot the scene, then re-shoot it at the same camera angles, this time you walking through the scene. Now take this to Video Editing Program, Open it up and depending which way the Camera is pointed, in my Corel VideoStudio10X, it points down from the top. so I would place the scene image on the top timeline, my text on the next timeline, then place the scene where you walk through it on the third timeline. This video track has to line up exactly lined up with the Video track above, especially if your using different camera angles. Also this is the Video track where you mask everything out of the scene except you. This will be a very time consuming Project to do, because you will have to mask every frame of the lower Video.
It can be done, I watched a Video on it not long ago, but I cannot remember where.
