POFMakeup: A style transfer method for Peking Opera makeup. (December 2022)

• Record Type: Journal Article
• Authors: Zhang, Fachao; Liang, Xiaoman; Sun, Yaqi; Lin, Mugang; Xiang, Jin; Zhao, Huihuang
• Is Part Of: Computers & electrical engineering. Volume 104:Part A(2022)
• Publication Date: 2022-12
• Subjects: Style transfer -- Face key point detection -- Facial segmentation -- Peking Opera; Computer engineering -- Periodicals; Electrical engineering -- Periodicals; Electrical engineering -- Data processing -- Periodicals; Ordinateurs -- Conception et construction -- Périodiques; Électrotechnique -- Périodiques; Électrotechnique -- Informatique -- Périodiques; Computer engineering; Electrical engineering; Electrical engineering -- Data processing; Periodicals; Electronic journals; 621.302854
• DOI: 10.1016/j.compeleceng.2022.108459
• Languages: English
• ISSNs: 0045-7906

Abstract

Highlights: A method for Peking Opera makeup based on one-example is proposed. The method does not require a GPU to participate in the computation, does not require a long training time (about 8 minutes for training on a CPU), and does not require a large dataset (only a few dozen images). The method ensures semantic consistency and preserves the identity of the target subject. A new face segmentation method is proposed, which can overcome the shortcomings of previous face segmentation using skin detection, and can segment faces quickly and accurately. An improved face key point calibration algorithm is proposed to realize the correct calibration of key points on the face of Peking Opera characters. Abstract: When using standard neural network transfer methods for portrait style transfer, semantically correct transfers often cannot be guaranteed; the texture details in the style examples tend to be ignored. This paper proposes a style transfer method for Peking Opera faces called POFMakeup. This method can transfer the style of a portrait with a Peking Opera face to another portrait. This method uses two guides, namely, a position guide and an appearance guide. The position guide ensures semantic consistency in the transfer process, while the appearance guide ensures the appearance of the target subject can be preserved during transfer. The experimental results show that this method not only solves the problem of traditional portrait style transfer but also can quickly and perfectly realize the style transfer of the Peking Opera face. The proposed method showed a 17% improvement in structural similarity (SSIM) over conventional methods. Graphical abstract: Image, graphical abstract