I don't propose to write lots about this. Instead I offer links to other sites who have spent time covering the subject
Our system uses the same principles: The two images from two projectors are superimposed in register on the screen, one for the left eye, and one for the right. One projector has a polarising filter which only allows light through in '/' direction, the other in the opposite '\' direction. Each participant in the lecture is provided with a pair of polarising spectacles - one lense allows '/' light in, the other allows '\' light in - like a seive! The result is the right and left eye each get their designated content - a minutely different view of the same scene. The only difference is the views were taken 40 to 60mm apart - the approximate spacing between our eyes. The result is we interpret this information as a 3-dimensional image. By placing a second Virtual Camera in 3D modelling software then, you can generate stereo images from models... of , for instance, anatomy.
For more detail, see: