There are programs which can combine multiple 2D photos to produce 3D models. But you'd need to take lots of photos from various angles, with consistent lighting and no distracting backgrounds. This could well get boring for the models; it's certainly not going to be a fast process with a single camera. Trying to use this technique to capture multiple people at the same time would be more difficult, and a lot slower.

I suppose you could record the positions of people with markers on them and replace them with posable models, if that was sufficient for your purpose; it would be faster. Why are you interested in recording postures anyway?