一种基于身份保持的多类面部表情图片生成技术

• Format: Patent
• Language: Chinese
• Published: 05.07.2024
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; PHYSICS

本发明涉及图像处理技术领域，特别涉及一种基于身份保持的多类面部表情图片生成技术，包括条件生成对抗网络模型训练、多类面部表情图片生成两个步骤。本发明能够根据主体中性表情图片生成包含该主体六类基本不同面部表情(愤怒、厌恶、恐惧、快乐、悲伤、以及惊讶)的面部图片，具有灵活性和实用性；另外，该技术能够在生成主体多类面部表情图片的同时保持主体身份信息，使得生成图片更为自然逼真，具有良好的应用前景。 The invention relates to the technical field of image processing, in particular to a multi-class facial expression picture generation technology based on identity preservation, which comprises two steps of conditional generative adversarial network model training and multi-class facial expression picture generation. According to the method, the facial picture containing six types of basic different facial expressions (anger, aversion, fear, happiness, sadness and surprising) of the subject can be generated according to the neutral expression picture of the subject, and the method has flexibility and practicability; besides, the technology can maintain identity information of the main body while generating various facial expression pictures of the main body, so that the generated pictures are more natural and vivid, a