Pengembangan Stetoskop Digital Untuk Klasifikasi Kelainan Katup Jantung Berbasis Deep Learning

Karimah, Izzah Khodijah (2025) Pengembangan Stetoskop Digital Untuk Klasifikasi Kelainan Katup Jantung Berbasis Deep Learning. Other thesis, Institut Teknologi Sepuluh Nopember.

[thumbnail of 5023211026-Undergraduate_Thesis.pdf] Text
5023211026-Undergraduate_Thesis.pdf - Accepted Version
Restricted to Repository staff only

Download (10MB) | Request a copy

Abstract

Valvular heart disease (VHD) adalah penyakit kardiovaskular dengan prevalensi tinggi di seluruh dunia, termasuk di Indonesia. Auskultasi, menggunakan stetoskop untuk mendengarkan suara jantung, adalah metode utama yang digunakan oleh profesional kesehatan dalam mendeteksi gangguan ini. Namun, hasil diagnostik metode ini sering menunjukkan variabilitas yang signifikan. Penelitian ini mengembangkan sistem klasifikasi otomatis kelainan katup jantung menggunakan sinyal Phonocardiogram (PCG) dan Photoplethysmography (PPG), memanfaatkan algoritma deep learning yang menggabungkan model Convolutional Neural Network (CNN) dan Long Short-Term Memory (LSTM). Tidak seperti sinyal dengan penanda morfologi yang jelas, sinyal PCG tidak memiliki morfologi tetap, sehingga sulit untuk mendeteksi suara jantung S1 dan S2 secara tepat. Untuk mengatasi ini, sinyal PPG digunakan sebagai referensi untuk segmentasi S1, memberikan informasi tambahan kepada model pembelajaran mesin untuk memperjelas posisi murmur. Sinyal PCG yang direkam di lingkungan nyata sering kali mengandung kebisingan dari suara lingkungan atau tubuh, yang dapat menurunkan kualitas analisis. Untuk menguranginya, proses reduksi kebisingan dilakukan pada sinyal PCG menggunakan Discrete Wavelet Transform (DWT) dengan wavelet Daubechies. Sistem ini mengintegrasikan metode Mel-Frequency Cepstral Coefficient (MFCC) sebagai ekstraksi fitur utama dari sinyal PCG untuk menangkap karakteristik spektral suara jantung. Pendekatan ini mencapai peningkatan akurasi klasifikasi untuk jenis defek katup, termasuk Aortic Stenosis (AS), Mitral Stenosis (MS), Mitral Regurgitation (MR), dan Mitral Valve Prolapse (MVP), dengan akurasi 93.33% dibandingkan 98.89% tanpa informasi posisi S1. Hasil penelitian ini adalah pengembangan perangkat stetoskop digital berbasis PCG-PPG yang dirancang secara ergonomis yang mampu memberikan diagnosis dini kelainan katup jantung dengan akurasi lebih tinggi. Sistem ini berpotensi menawarkan solusi praktis dalam diagnosis dan deteksi dini VHD, terutama di daerah dengan akses terbatas ke teknologi canggih dan tenaga ahli.
=================================================================================================================================
Valvular heart disease (VHD) is a cardiovascular disease with high global prevalence, including in Indonesia. Auscultation, using a stethoscope to listen to heart sounds, is the primary method for detecting this disorder by health professionals. However, this method's diagnostic results often exhibit significant variability. This study develops an automatic classification system for heart valve abnormalities using Phonocardiogram (PCG) and Photoplethysmography (PPG) signals, leveraging a deep learning algorithm combining Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM) models. Unlike signals with clearly defined morphological landmarks, PCG signals lack a consistent morphology, making it challenging to precisely detect the S1 and S2 heart sounds. To address this, the PPG signal is utilized as a reference for S1 segmentation, providing additional information to the machine learning model to clarify the murmur's position. PCG signals recorded in real environments often contain noise from environmental or body sounds, which can degrade analysis quality. To mitigate this, noise reduction is performed on PCG signals using Discrete Wavelet Transform (DWT) with Daubechies wavelets. The system incorporates the Mel-Frequency Cepstral Coefficient (MFCC) method as the main feature extraction technique from PCG signals to capture the spectral characteristics of heart sounds. This approach achieved improved classification accuracy for valve defect types, including Aortic Stenosis (AS), Mitral Stenosis (MS), Mitral Regurgitation (MR), and Mitral Valve Prolapse (MVP), with an accuracy of 93.33% compared to 98.89% without S1 positional information. The outcome of this research is the development of an ergonomically designed PCG-PPG-based digital stethoscope device that provides early diagnosis of heart valve abnormalities with higher accuracy. This system has the potential to offer practical solutions in VHD diagnosis and early detection, especially in areas with limited access to advanced technology and experts.

Item Type: Thesis (Other)
Uncontrolled Keywords: Valvular Heart Disease, Phonocardiogram, Photoplethysmography, MFCC, Deep Learning, CNN-LSTM, Valvular Heart Disease, Phonocardiogram, Photoplethysmography, MFCC, Deep Learning, CNN-LSTM
Subjects: R Medicine > R Medicine (General) > R856.2 Medical instruments and apparatus.
R Medicine > R Medicine (General) > R858 Deep Learning
R Medicine > RC Internal medicine > RC78 Diagnosis, Radioscopic--Examinations, questions, etc.
T Technology > T Technology (General) > T57.8 Nonlinear programming. Support vector machine. Wavelets. Hidden Markov models.
T Technology > T Technology (General) > T58.62 Decision support systems
T Technology > TA Engineering (General). Civil engineering (General) > TA1573 Detectors. Sensors
T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK5102.9 Signal processing.
Divisions: Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Biomedical Engineering > 11410-(S1) Undergraduate Thesis
Depositing User: Izzah Khodijah Karimah
Date Deposited: 04 Aug 2025 01:08
Last Modified: 04 Aug 2025 01:08
URI: http://repository.its.ac.id/id/eprint/126610

Actions (login required)

View Item View Item