| dc.description.abstract | Acne on facial skin is one of the most common dermatological problems, with a high prevalence across various age groups. This condition not only affects skin health but also has a significant impact on the psychological well-being of individuals. Manual identification of acne is often inconsistent, highlighting the need for an automated system capable of accurately detecting and classifying acne types. This study aims to develop a system for detecting and classifying mild to moderate facial acne using the You Only Look Once version 11 (YOLOv11) algorithm, while also providing lifestyle recommendations based on acne location through the Decision Tree method. The dataset was obtained from DermNet, ACNE04 Dataset, and Roboflow covering five acne types: blackheads, whiteheads, papules, pustules, and nodules. The research process included preprocessing, data augmentation, YOLOv11 model training, and the integration of Decision Tree rules to generate lifestyle recommendations. Experimental results show that the YOLOv11 model achieved a Mean Average Precision (mAP) of 91%, precision of 92.2%, recall of 82.4%, and an F1-score of 85%. Testing the Decision Tree system with 70 facial images yielded an accuracy of 85.8%, precision of 98.7%, recall of 86.8%, and an F1-score of 92.4%. The system was implemented as a web-based application and tested on 35 respondents of different ages, genders, and occupations. User testing demonstrated positive results in terms of interface design, ease of use, and the relevance of acne detection and lifestyle recommendations. Overall, this study developed a computer vision and machine learning-based system that not only detects and classifies acne but also provides practical lifestyle recommendations to support daily facial skin care. | en_US |
