This thesis focus on human action recognition using volumetric reconstructions obtained from multiple monocular cameras. The problem of action recognition has been addressed using di erent approaches, both in the 2D and ...