Journal of Computer Applications ›› 2023, Vol. 43 ›› Issue (2): 601-607.DOI: 10.11772/j.issn.1001-9081.2022010030

• Multimedia computing and computer simulation • Previous Articles    

Controllable face editing algorithm with closed-form solution

Lingling TAO1,2, Bo LIU1,2(), Wenbo LI1,2, Xiping HE1,2   

  1. 1.School of Artificial Intelligence,Chongqing Technology and Business University,Chongqing 400067,China
    2.Chongqing Key Laboratory of Intelligent Perception and BlockChain Technology (Chongqing Technology and Business University),Chongqing 400067,China
  • Received:2022-01-11 Revised:2022-04-05 Accepted:2022-04-11 Online:2022-04-28 Published:2023-02-10
  • Contact: Bo LIU
  • About author:TAO Lingling, born in 1998, M. S. candidate. Her research interests include computer vision, image processing, generative adversarial network.
    LI Wenbo, born in 1998, M. S. candidate. His research interests include machine learning, computer vision, image generation.
    HE Xiping, born in 1968, Ph. D., professor. His research interests include machine learning, data analysis and processing, computer vision.
  • Supported by:
    Key Platform Fund of Chongqing Technology and Business University(950119093);Graduate Innovation Project of Chongqing Technology and Business University(yjscxx2021-112-99)


陶玲玲1,2, 刘波1,2(), 李文博1,2, 何希平1,2   

  1. 1.重庆工商大学 人工智能学院,重庆 400067
    2.智能感知与区块链技术重庆市重点实验室(重庆工商大学),重庆 400067
  • 通讯作者: 刘波
  • 作者简介:陶玲玲(1998—),女,重庆人,硕士研究生,主要研究方向:计算机视觉、图像处理、生成对抗网络
  • 基金资助:


To solve the problems in face editing, such as unnatural editing results and great changes in generated images, a controllable face editing algorithm with closed-form solution was proposed. Firstly, n latent vectors were sampled randomly to construct a sample matrix, and the top k principal component vectors of the matrix were calculated. Then, five attributes of face image were obtained by ResNet-50, and the semantic boundary of each attribute was calculated by Support Vector Machine (SVM). Finally, the interpretable direction vectors of these attributes were calculated, which were as closed to the principal components vectors as possible and stayed as far away from the semantic boundary of the corresponding attribute as possible at the same time, thereby reducing the coupling between facial attributes, and improving the controllability in face editing. Because the algorithm has a closed-form solution, it has high efficiency. Experimental results show that the compared with closed-form Factorization of latent Semantics in GANs (SeFa) algorithm and Discovering Interpretable Generative Adversarial Network Controls (GANSpace) algorithm, the proposed algorithm increases the Inception Score (IS) by 19% and 26% respectively, decreases the Fréchet Inception Distance (FID) by 4% and 37% respectively, and decreases the Maximum Mean Discrepancy (MMD) by 15% and 48% respectively. It can be seen that this algorithm has good controllability and decoupling.

Key words: Generative Adversarial Network (GAN), face editing, latent space, semantic space, attribute semantic boundary



关键词: 生成对抗网络, 人脸编辑, 潜在空间, 语义空间, 属性语义边界

CLC Number: