Graph Matching for Marker Labeling and Missing Marker Reconstruction With Bone Constraint by LSTM in Optical Motion Capture

Optical motion capture (MOCAP) is a commonly used technology to record the of non-rigid objects with high accuracy in 3D space. However, MOCAP data has be processed further before it can used. The scattered reconstructed must constitute human configuration by labelling process according predefined template, and missing markers have produce stable trajectory. In this work, we propose novel method for sequences. First, graph matching employed determine connection relationship single frame. Then, Kalman filtering tracking sequence. As challenge coming from markers, new preprocessing considering bone length constraint, which represents information variation relative position adjacent markers. input into Long-Short Term Memory (LSTM) model recover de-noise data. experiment conducted on our own dataset proves that achieves similar effect Cortex, commercial analysis software. CMU demonstrates marker reconstruction achieve an art-of-state result. code will pulished https://github.com/Lijianfang6930/Graph-Matching-for-Marker-Labelling