This might be possible for a somewhat skilled person. Basically, if you subtract the camera move, or at least limit it to a pan this effect becomes much easier. Shoot your frozen people in front of a green screen and insert them into the composition in proper perspective in the scene. You will then either have to shoot your walking guy in front of the green screen as well or do a lot of roto work. Once you have the elements put them in front and behind your main guy to make them look 3D. If you panned you will have to motion track the footage so that people don’t float or slide around. Basically, unless you nail the lighting, keying, tracking, and compositing this won’t be very impressive.
Another alternative is to have people stand still and keep the camera moving fast enough to avoid noticing this (usually obvious) fact. You might even combine the two.
Bottom line, if you’re not a pro with lots of time and money, try something simpler.