Pengembangan Stetoskop Digital untuk Deteksi Penyakit Paru Obstruktif Kronis (PPOK) Berbasis Ekstraksi Fitur Mel-Frequency Cepstral Coefficient (MFCC)

Nadhilah, Rihhadatul Aisy (2025) Pengembangan Stetoskop Digital untuk Deteksi Penyakit Paru Obstruktif Kronis (PPOK) Berbasis Ekstraksi Fitur Mel-Frequency Cepstral Coefficient (MFCC). Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5023211020-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (5MB) | Request a copy

Abstract

Penyakit Paru Obstruktif Kronis (PPOK) merupakan kondisi paru-paru dimana saluran udara yang masuk dan keluar menjadi tersumbat serta adanya penurunan fungsi paru-paru yang progresif. Gejala PPOK meliputi batuk kronis, yang terkadang disertai dahak, kesulitan bernapas, mengi pada suara napas, dan kelelahan. Kondisi yang terus menerus memburuk dapat menyebabkan penurunan fungsi paru-paru serta komplikasi penyakit lainnya pada pasien. Untuk mencegah terjadinya hal tersebut, dibutuhkan teknologi deteksi dini PPOK agar pengobatan dapat diberikan secara cepat dan tepat bagi pasien. Sistem deteksi PPOK ini mengintegrasikan sensor MAX4466 dengan stetoskop dan rangkaian instrumentasi filter analog aktif, rangkaian ¬non-inverting adder amplifier, dan clipper. Mikrokontroler STM32L031K6 yang digunakan berhasil mengakuisisi data sinyal suara paru-paru yang kemudian dikirimkan ke Raspberry Pi melalui komunikasi UART USB. Sistem pengolahan sinyal menggunakan metode pemrosesan Discrete Wavelet Transform (DWT) dengan ekstraksi fitur Mel-Frequency Cepstral Coefficients (MFCC). Kombinasi kedua metode tersebut dapat memisahkan sinyal suara paru-paru dengan sinyal lain yang terekam dan merepresentasikan karakteristik dan pola sinyal suara paru-paru dengan baik. Model klasifikasi yang digunakan adalah Bidirectional Long-Short Time Memory (BiLSTM) untuk memprediksi sinyal suara paru-paru ke dalam kelas Normal dan PPOK. Penelitian ini menghasilkan performa model klasifikasi yang sangat baik untuk tipe dataset yang sangat beragam. Dimana nilai akurasi pelatihan model adalah 91.06%, akurasi validasi model adalah 89.05%, dan akurasi pengujian model adalah 89.34%. Kemudian, model dapat dianggap seimbang dan tidak bias ke salah satu kelas yang ditunjukkan dengan nilai sensitivitas dan nilai spesifisitas model yaitu 89%. Hal ini menunjukkan model mampu mendeteksi PPOK dengan kesalahan yang minimal. Hasil penelitian menunjukkan bahwa stetoskop digital yang dikembangkan dapat secara efektif mendeteksi PPOK, dan memberikan kontribusi dalam bidang ilmu pengetahuan dan kesehatan penyakit pernapasan.
========================================================================================================================================
Chronic Obstructive Pulmonary Disease (COPD) is a lung condition characterized by airway obstruction and progressive decline in lung function. COPD symptoms include chronic cough, sometimes accompanied by phlegm, difficulty breathing, wheezing, and fatigue. Persistent worsening of the condition can result in further lung function decline and additional health complications for patients. To prevent such outcomes, early detection technology for COPD is essential to ensure prompt and accurate treatment for patients. The COPD detection system in this study integrates the MAX4466 sensor with a stethoscope and an analog active filter instrumentation circuit, a non-inverting adder amplifier, and a clipper circuit. The STM32L031K6 microcontroller was successfully used to acquire lung sound signals, which were then transmitted to a Raspberry Pi via USB UART communication. The signal processing pipeline employed the Discrete Wavelet Transform (DWT) method combined with Mel-Frequency Cepstral Coefficients (MFCC) feature extraction. This combination effectively separates lung sound signals from other recorded noises and represents the acoustic characteristics and patterns of lung sounds with high fidelity. The classification model used is a Bidirectional Long Short-Term Memory (BiLSTM) network, which classifies the lung sounds into Normal and COPD classes. The model demonstrated excellent classification performance across a highly diverse dataset, achieving 91.06% training accuracy, 89.05% validation accuracy, and 89.34% testing accuracy. Moreover, the model is considered balanced and unbiased, as indicated by both its sensitivity and specificity scores of 89%. These results highlight the model’s ability to detect COPD with minimal error. Overall, the study shows that the developed digital stethoscope system can effectively detect COPD and contributes meaningfully to advancements in respiratory health technology and clinical diagnostics.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Kata kunci: PPOK, suara paru-paru, stetoskop digital, DWT, MFCC, BiLSTM Keywords: COPD, lung sounds, digital stethoscope, DWT, MFCC, BiLSTM.
Subjects:	T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK5102.9 Signal processing. T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK7871.674 Detectors. Sensors T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK7872.F5 Filters (Electric) T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK7878 Electronic instruments
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Biomedical Engineering > 11410-(S1) Undergraduate Thesis
Depositing User:	Rihhadatul Aisy Nadhilah
Date Deposited:	04 Aug 2025 02:39
Last Modified:	04 Aug 2025 02:39
URI:	http://repository.its.ac.id/id/eprint/125696

Actions (login required)

View Item