Penerapan Fluoresensi Optik dan Pengukuran Konduktivitas Pada Saliva untuk Deteksi Dini Karies Gigi Menggunakan Metode Neural Network

Yusril, Muhammad (2025) Penerapan Fluoresensi Optik dan Pengukuran Konduktivitas Pada Saliva untuk Deteksi Dini Karies Gigi Menggunakan Metode Neural Network. Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 6022231006-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (7MB) | Request a copy

Abstract

Karies gigi merupakan salah satu penyakit dengan prevalensi tertinggi secara global dan menjadi isu signifikan dalam bidang kesehatan masyarakat, khususnya di Indonesia. Deteksi dini terhadap karies memiliki peran krusial dalam mencegah kerusakan struktural lebih lanjut, namun metode konvensional seperti inspeksi visual dan radiografi masih menghadapi keterbatasan dalam hal sensitivitas, subjektivitas pemeriksa, serta keterjangkauan teknologi. Penelitian ini mengusulkan suatu sistem biosensor non-invasif berbasis saliva untuk deteksi dini karies gigi, yang mengintegrasikan analisis fluoresensi optik dan pengukuran konduktivitas elektrokimia, serta diproses menggunakan algoritma neural network. Kurkumin digunakan sebagai fluorofor alami yang intensitas fluoresensinya menurun secara spesifik ketika berinteraksi dengan enzim MMP-8, suatu biomarkerdegradasi matriks kolagen pada jaringan gigi. Emisi cahaya hasil eksitasi oleh sinar UV pada 365 nm diukur menggunakan sensor spektral AS7262, sedangkan parameter konduktivitas saliva diperoleh melalui teknik Linear Sweep Voltammetry (LSV) dengan elektroda karbon cetak (SPCE). Seluruh data yang diperoleh diolah menggunakan model Artificial Neural Network (ANN) dan Convolutional Neural Network satu dimensi (CNN 1D) untuk mengklasifikasikan kondisi saliva menjadi tiga kategori: sehat, karies tahap awal, dan karies lanjut. Hasil penelitian menunjukkan bahwa kombinasi data fluoresensi dan konduktivitas yang dianalisis menggunakan CNN 1D menghasilkan akurasi klasifikasi tertinggi sebesar 85,71%, menunjukkan kapabilitas tinggi model dalam mengenali pola non-linear dari data biologis kompleks. Sistem ini menunjukkan potensi sebagai alat deteksi dini karies gigi yang non-invasif, dan portabel, serta membuka peluang pengembangan perangkat skrining mandiri berbasis kecerdasan buatan.
===================================================================================================================================
Dental caries is one of the most prevalent diseases globally and represents a significant public health issue, particularly in Indonesia. Early detection of caries plays a crucial role in preventing further structural damage; however, conventional methods such as visual inspection and radiography still face limitations in terms of sensitivity, examiner subjectivity, and technological accessibility. This study proposes a non-invasive saliva-based biosensor system for the early detection of dental caries, integrating optical fluorescence analysis and electrochemical conductivity measurement, processed using neural network algorithms. Curcumin is used as a natural fluorophore whose fluorescence intensity specifically decreases upon interaction with MMP-8 enzyme, a biomarker for collagen matrix degradation in dental tissue. The emitted light resulting from UV excitation at 365 nm is measured using an AS7262 spectral sensor, while saliva conductivity parameters are obtained through Linear Sweep Voltammetry (LSV) using a screen-printed carbon electrode (SPCE). All acquired data are processed using Artificial Neural Network (ANN) and one-dimensional Convolutional Neural Network (1D CNN) models to classify saliva conditions into three categories: healthy, early-stage caries, and advanced caries. The results show that the combination of fluorescence and conductivity data analyzed with the 1D CNN model achieved the highest classification accuracy of 85.71%, demonstrating the model’s strong capability to recognize non-linear patterns in complex biological data. This system shows promising potential as a non-invasive, portable early detection tool for dental caries and opens opportunities for the development of AI-based self-screening devices.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	Karies Gigi, Saliva, MMP-8, Kurkumin, Fluoresensi, Konduktivitas, Neural Network, Sensor Spektrum AS7262 Dental Caries, Saliva, MMP-8, Curcumin, Fluorescence, Conductivity, Neural Network, AS7262 Spectral Sensor
Subjects:	T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK7871.674 Detectors. Sensors
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Electrical Engineering > 20101-(S2) Master Thesis
Depositing User:	Muhammad Yusril
Date Deposited:	29 Jul 2025 10:02
Last Modified:	29 Jul 2025 10:02
URI:	http://repository.its.ac.id/id/eprint/122352

Actions (login required)

View Item