Implementasi Deep Learning dalam klasifikasi Penyakit Kulit Wajah Berbasis Convolutional Neural Network dan Transfer Learning

Yasmin, Dini Athirah (2025) Implementasi Deep Learning dalam klasifikasi Penyakit Kulit Wajah Berbasis Convolutional Neural Network dan Transfer Learning. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5003211033-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (9MB) | Request a copy

Abstract

Kulit wajah merupakan bagian tubuh yang paling terekspos terhadap berbagai faktor lingkungan seperti paparan sinar matahari, polusi, dan kelembapan yang menjadikannya rentan terhadap berbagai gangguan kesehatan. Indonesia sebagai negara tropis dengan tingkat kelembapan udara dan paparan sinar matahari yang tinggi menjadikan kulit wajah rentan terhadap gangguan seperti iritasi, peradangan, hiperpigmentasi, dermatitis atopik, hingga jerawat. Diagnosis dan pengobatan penyakit kulit secara cepat dan akurat sangat penting untuk mencegah komplikasi dan mempercepat proses penyembuhan. Namun, keterbatasan akses terhadap layanan dermatologis, terutama di daerah terpencil, menjadi kendala tersendiri. Dalam menjawab tantangan tersebut, penelitian ini mengembangkan sistem klasifikasi otomatis berbasis citra menggunakan algoritma Convolutional Neural Network (CNN) dan transfer learning untuk mendeteksi delapan jenis penyakit kulit wajah, yaitu melasma, flek hitam, herpes, jerawat, Rosacea, milia, panu, dan tratak (pytariasis alba). Dataset yang digunakan terdiri dari 893 gambar yang diperoleh dari proses scraping di DermNet dan Bing Images. Tiga arsitektur model diterapkan dan dibandingkan, yaitu CNN konvensional, MobileNetV2, dan DenseNet121. Berdasarkan hasil evaluasi, model DenseNet121 dengan penerapan augmentasi data dan fine tuning menunjukkan performa terbaik dengan akurasi validasi sebesar 91,01%, loss validasi sebesar 0,4338, dan akurasi pengujian sebesar 87%. Temuan ini menunjukkan bahwa pendekatan yang diusulkan mampu menghasilkan sistem klasifikasi yang akurat dan potensial untuk mendukung proses diagnosis awal penyakit kulit wajah.
====================================================================================================================================
Facial skin is the part of the body that is most exposed to various environmental factors such as sun exposure, pollution, and moisture which makes it vulnerable to various health problems. Indonesia as a tropical country with high humidity and sun exposure makes facial skin prone to disorders such as irritation, inflammation, hyperpigmentation, atopic dermatitis, and acne. Quick and accurate diagnosis and treatment of skin diseases is essential to prevent complications and speed up the healing process. However, limited access to dermatological services, especially in remote areas, is an obstacle. In response to these challenges, this research develops an image-based automatic classification system using Convolutional Neural Network (CNN) and transfer learning algorithms to detect eight types of facial skin diseases, namely melasma, dark spots, herpes, acne, Rosacea, milia, tinea versicolor, and tratak (pytariasis alba). The dataset used consists of 893 images obtained from scraping on DermNet and Bingg Images. Three model architectures were applied and compared, namely conventional CNN, MobileNetV2, and DenseNet121. Based on the evaluation results, the DenseNet121 model with the application of augmentation and fine tuning showed the best performance with a validation accuracy of 91,01%, validation loss of 0,4338, and testing accuracy of 87%. These findings indicate that the proposed approach is capable of producing an accurate classification system and has the potential to support the early diagnosis process of facial skin diseases.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Convolutional Neural Network, Penyakit Kulit Wajah, Transfer Learning, Convolutional Neural Network, Facial Skin Disease, Transfer Learning
Subjects:	H Social Sciences > HD Industries. Land use. Labor > HD108 Classification (Theory. Method. Relation to other subjects ) Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines. Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science) Q Science > QA Mathematics > QA76.9.D343 Data mining. Querying (Computer science) R Medicine > R Medicine (General) > R858 Deep Learning
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49201-(S1) Undergraduate Thesis
Depositing User:	Dini Athirah Yasmin
Date Deposited:	05 Aug 2025 02:14
Last Modified:	05 Aug 2025 02:14
URI:	http://repository.its.ac.id/id/eprint/127285

Actions (login required)

View Item