US 12,270,910 B2
System and method of capturing three-dimensional human motion capture with LiDAR
Cheng Wang, Xiamen (CN); Jialian Li, Xiamen (CN); Lan Xu, Shanghai (CN); Chenglu Wen, Xiamen (CN); and Jingyi Yu, Shanghai (CN)
Assigned to Xiamen University, Xiamen (CN); and ShanghaiTech University, Shanghai (CN)
Filed by XIAMEN UNIVERSITY, Fujian (CN); and SHANGHAITECH UNIVERSITY, Shanghai (CN)
Filed on Aug. 9, 2022, as Appl. No. 17/884,273.
Application 17/884,273 is a continuation of application No. PCT/CN2022/078083, filed on Feb. 25, 2022.
Prior Publication US 2023/0273318 A1, Aug. 31, 2023
Int. Cl. G01S 17/89 (2020.01); G01S 17/86 (2020.01); G06N 20/00 (2019.01)
CPC G01S 17/89 (2013.01) [G01S 17/86 (2020.01); G06N 20/00 (2019.01)] 31 Claims
OG exemplary drawing
 
1. A computer-implemented method for training machine learning models to generate three-dimensional motions based on LiDAR point clouds, the method comprising:
encoding, by a computing system, a machine learning model representing an object in a scene; and
training, by the computing system, the machine learning model using a dataset comprising synchronous LiDAR point clouds captured by monocular LiDAR sensors and ground-truth three-dimensional motions obtained from Inertial Measurement Units (IMU) devices;
wherein the machine learning model is configured to generate a three-dimensional motion of the object based on an input of a plurality of point cloud frames captured by a monocular LiDAR sensor, and
the machine learning model comprises a feature learning network, a Gated Recurring Unit (GRU), a multiplayer perceptron (MLP) decoder, wherein the feature learning network is configured to extract a global descriptor from each point cloud frame, the GRU is configured to generate a plurality of hidden variables for the global descriptor, and the MLP decoder is configured to predict a plurality of joint locations based on the plurality of hidden variables.