Fig. 7From: Hands-on: deformable pose and motion models for spatiotemporal localization of fine-grained dyadic interactionsMean AuC over all interactions using four-fold cross-validation in the SC (left) and MC (right) scenarios of ShakeFive2Back to article page