Method and system for intelligent generation of three-dimensional digital content based on motion capture

Abstract

The application relates to the technical field of three-dimensional digital content generation, and discloses a three-dimensional digital content intelligent generation method and system based on motion capture, which comprises the following steps: inputting motion capture data of a multi-person interactive segment into a rule knowledge base for frame-by-frame matching detection, determining the compliance of each type of interactive rule one by one, and marking all abnormal time points and corresponding abnormal skeleton nodes of roles that violate social interaction rules; the application pre-constructs a social interaction rule knowledge base containing a human distance threshold rule, a gaze matching rule, a body orientation coordination rule and a speaking turn response rule, carries out frame-by-frame matching detection on the social interaction logic level of multi-person motion capture data, achieves the effect of automatically identifying semantic deviations in the original motion capture data that do not conform to social common sense, and simultaneously accurately locks the multi-person interactive segment and skips the single-person independent action segment through a time sequence segmentation mechanism based on voice activity and body relative position change, so that the invalid calculation amount is reduced.

Images