What 3D scene reconstruction methods does the term "multi-view stereo" encompass? Is it only used for methods that apply binocular stereo algorithms (taking 2 views as input) in a pairwise manner? Or only methods based on disparity estimation? If not why is the term "stereo" used?
Multi-view stereo (MVS) is the general term given to a group of techniques that use stereo correspondence as their main cue and use more than two images.
A quote from 'Multi-View Stereo: A Tutorial' by Yasutaka Furukawa and Carlos Hernández.
So to paraphrase: We have a set of images that is larger than two, and use them in a pairwise manner by applying techniques that finde stereo correspondences to reconstruct the object shown in them.
If you want to know more about MVS algorithms that paper I linked seems to be pretty good and at least Furukawa is a name that I know and have read papers from, so he seems to know the topic.