Fig. 9From: Hands-on: deformable pose and motion models for spatiotemporal localization of fine-grained dyadic interactionsMean AuC over all interactions using four-fold cross-validation with different numbers of training sequences, for SC (left) and MC (right)Back to article page