• Journal of Internet Computing and Services
    ISSN 2287 - 1136 (Online) / ISSN 1598 - 0170 (Print)
    https://jics.or.kr/
Menu

Efficient Facial Feature Classification for Makeup-Robust Recognition (EFFC)


Su-Rak Son, Yi-Na Jeong, Journal of Internet Computing and Services, Vol. 26, No. 6, pp. 43-50, Dec. 2025
10.7472/jksii.2025.26.6.43, Full Text:  HTML
Keywords: Deep Learning, CNN, Facial Recognition, facial feature classification

Abstract

This paper proposes EFFC (Efficient Facial Feature Classification) for makeup-robust facial recognition. In the preprocessing stage, 68 facial landmarks are used to segment and align key regions (eyes, nose, mouth), which are then normalized into shape labels (e.g., eye-tail direction, nasal bridge height, lip thickness). A lightweight CNN learns each region without hand-crafted rules, reducing dependence on color/texture cues and emphasizing geometric stability. On a synthetic face set, EFFC achieved high accuracies for several regions (eye-tail 95.57%, face shape 93.18%) with an average improvement of over 2 percentage points. The compact design enables real-time and embedded deployment. Future work includes validation on real makeup/no-makeup pairs and cross-dataset evaluations.


Statistics
Show / Hide Statistics

Statistics (Past 3 Years)
Multiple requests among the same browser session are counted as one view.
If you mouse over a chart, the values of data points will be shown.


Cite this article
[APA Style]
Son, S. & Jeong, Y. (2025). Efficient Facial Feature Classification for Makeup-Robust Recognition (EFFC). Journal of Internet Computing and Services, 26(6), 43-50. DOI: 10.7472/jksii.2025.26.6.43.

[IEEE Style]
S. Son and Y. Jeong, "Efficient Facial Feature Classification for Makeup-Robust Recognition (EFFC)," Journal of Internet Computing and Services, vol. 26, no. 6, pp. 43-50, 2025. DOI: 10.7472/jksii.2025.26.6.43.

[ACM Style]
Su-Rak Son and Yi-Na Jeong. 2025. Efficient Facial Feature Classification for Makeup-Robust Recognition (EFFC). Journal of Internet Computing and Services, 26, 6, (2025), 43-50. DOI: 10.7472/jksii.2025.26.6.43.