Personalized shadow puppet face generation method based on constraint generative adversarial network

The invention discloses a personalized shadow puppet face generation method based on a constraint generative adversarial network, and the method comprises the steps: building a generative adversarial network which is used for converting a real face image into a shadow puppet face image, and introduc...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: HUANG JINGZHOU, ZHOU HOUPAN, JIANG JINGCHAO
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 personalized shadow puppet face generation method based on a constraint generative adversarial network, and the method comprises the steps: building a generative adversarial network which is used for converting a real face image into a shadow puppet face image, and introducing a local geometric feature constraint function into the generative adversarial network, calculating an image domain required by loss generation through a local geometric feature constraint function, and obtaining a generative adversarial network model meeting requirements by constraining geometric deformation of typical facial features in a training process of the generative adversarial network; and inputting a real face image into a generator of the generative adversarial network model, and generating a shadow puppet face image of which part of features are consistent with the real face. According to the method, the defect that the shadow puppet face generated due to texture loss and geometric structure transfo