Radhiyatammardhiyyah
Politeknik Negeri lhokseumawe

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search

Facial Skin Color Classification Using Hybrid Features With The K-Nearest Neighbor (KNN) Method Nurul Agustina; Radhiyatammardhiyyah; Cut Aulia Safira Rahman
International Journal of Applied Artificial Intelligence and Robotics Vol. 2 No. 2 (2026): Vol 2 No 2 (November 2026)
Publisher : PT. Literasi Teknologi Informatika

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.67745/ijaic.v2i2.26

Abstract

Facial skin color classification can still be affected by differences in lighting, cameras, and image conditions, causing visual assessment to potentially produce inconsistent results. This study aims to develop a facial skin color classification system using a combination of color and texture features and the K-Nearest Neighbor (KNN) algorithm. The research data consist of primary and secondary datasets. The research stages include data collection, preprocessing, color feature extraction using Color Moments in the YCbCr color space, texture feature extraction using Gray Level Co-occurrence Matrix (GLCM), hybrid feature construction, KNN classification, and performance evaluation. The color and texture features are combined into a hybrid feature representation, which is then used as the input for KNN. K-value testing was conducted using several configurations to determine the optimal parameter. The results show that K = 3 achieved the highest performance in both dataset scenarios. The primary dataset achieved an accuracy of 90.61%, while the secondary dataset achieved an accuracy of 87.43%. Class-wise evaluation indicates that the model can recognize all three skin color categories with good performance. The results demonstrate that combining color and texture information can be an effective approach for facial skin color classification.