Three-dimensional human body reconstruction method and system based on STAR model

The invention discloses a three-dimensional human body reconstruction method and system based on an STAR model, and the method comprises the steps: carrying out the feature extraction of input human body information through employing a convolutional neural network, obtaining the global features of a...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: CONG XUNCHAO, GAN YI
Format: Patent
Sprache:chi ; eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:The invention discloses a three-dimensional human body reconstruction method and system based on an STAR model, and the method comprises the steps: carrying out the feature extraction of input human body information through employing a convolutional neural network, obtaining the global features of a human body, inputting the global features into a time sequence feature extraction network, carrying out the coding of time-related features in a video, and carrying out the coding of the time-related features; a group of feature sequences containing time sequence information is obtained, the human body feature sequences are used, a full-connection layer is used as a human body parameter regression network to predict posture parameters and body shape parameters of the SMPL, and a self-attention mechanism is added to improve the stability of an output human body model sequence. A posture parameter and a body shape parameter of the STAR model are predicted through the human body model parameter regression network, a