Title: Beauty aids: can AI improve human behaviours with imperfect data?

Authors: Wen-Feng Wang; Bai-Zhou Xu; Bin Hu; Fuqing Li; Lalit Mohan Patnaik; Lu-Jie Cui; Yun-Zhu Pan

Addresses: Faculty of Intelligent Technology, Shanghai Institute of Technology, Shanghai, 201418, China ' School of Computer Science and Technology, Hainan University, Haikou, 570228, China ' School of Information Science and Engineering, Changsha Normal University, Changsha, 410111, China ' School of Marxism, Xiangtan University, Xiangtan, 411100, China ' National Institute of Advanced Studies, Indian Institute of Science Campus, Bangalore, 560012, India ' IEDA Research Institute, Shanghai Institute of Technology, Shanghai 201418, China ' IEDA Research Institute, Shanghai Institute of Technology, Shanghai 201418, China

Abstract: This paper aims to examine whether AI can improve human behaviours with imperfect data. Beauty aids with the pretrained AI model is taken as a practical example. This model integrated fuzzy reasoning with ResNet-50 for facial beauty prediction (FBP) and real-time recommendations of makeup behaviours. Results shown that the AI model can provide beauty aids for people whose facial data have not be included during the pretraining process and improve their makeup behaviours. The difference between the maximal and minimal values amounts to 33.62, implying that the effect of beauty aids is evident. The cross validation with perfect data further also confirmed that the effects of increased makeup experiences are worthy of further attention. The recommended degree of powder makeup for the volunteer is 0.118~0.2, while that of lipstick and blush makeup is 0.034~0.2. As an emerging technique, potential evolutions of the real-time beauty aids system with AI and data science will bring out the long-term future of FBP research.

Keywords: models; parameters; data collection; makeup behaviours; FBP; facial beauty prediction.

DOI: 10.1504/IJDS.2025.149851

International Journal of Data Science, 2025 Vol.10 No.2, pp.119 - 135

Received: 08 May 2024
Accepted: 08 Nov 2024
Published online: 14 Nov 2025 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article