The invention provides a digital person display method and
system based on an intelligent VR space, and relates to the technical field of three-dimensional models.The method comprises the steps that multi-dimensional
feature extraction is conducted on an original person video, and a
dynamic feature data set is obtained;
joint point cloud construction and registration are carried out based on the
dynamic feature data set, and a space-time consistency skeleton model is obtained; performing surface
mask generation on the video frame of the original character video through the space-time consistency skeleton model to obtain a dynamic
human body grid with a separated background; and carrying out physical illumination fusion on the dynamic
human body grid and the VR environment parameters to obtain a
digital human VR
entity model. According to the method,
dynamic feature data set construction,
joint point cloud registration, parameterized
human body grid generation and physical illumination fusion are combined, high-fidelity
digital human migration from a common video to a VR space is achieved, the key problems of motion
distortion, rendering penetration and the like are solved, the motion track error is reduced, and the human
body contour segmentation precision is improved.