Generation method and device for mixed attack style face picture

• Main Authors: SU ANYANG, TANG DARUN, LIU XING, ZHAO CHENXU
• Format: Patent
• Language: Chinese; English
• Published: 24.10.2023
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; PHYSICS

The invention relates to the technical field of face recognition, and discloses a mixed attack style face picture generation method and device, and the method comprises the steps: carrying out the inverse mapping of a face picture sample containing a plurality of attack styles into a corresponding potential coding sample through a generator obtained through pre-training; carrying out attack style mixing on the potential coding sample to obtain a potential code of a mixed attack style; and inputting the potential code of the mixed attack style into the generator to obtain a target face picture with the mixed attack style output by the generator. 本申请涉及人脸识别技术领域，公开一种用于混合攻击风格人脸图片的生成方法及设备，其中，所述生成方法包括：通过预训练得到的生成器，将包含多种攻击风格的人脸图片样本逆映射为对应的潜在编码样本；对所述潜在编码样本进行攻击风格混合，得到混合攻击风格的潜在编码；将所述混合攻击风格的潜在编码输入至所述生成器中，得到所述生成器输出的具有混合攻击风格的目标人脸图片。