Deteksi dan Klasifikasi Laju Pernapasan Secara Non-Kontak Menggunakan Infrared Thermography

Salsabila, Anadya Ghina (2024) Deteksi dan Klasifikasi Laju Pernapasan Secara Non-Kontak Menggunakan Infrared Thermography. Diploma thesis, Institut Teknologi Sepuluh Nopember.

Text 5023201033-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2026. Download (4MB) | Request a copy

Abstract

Pemantauan parameter fisiologis vital seperti respiration rate (RR) merupakan indikator penting dalam menilai kondisi kesehatan pasien. Namun, teknik pengukuran yang ada saat ini umumnya memerlukan pemasangan sensor pada tubuh pasien, yang seringkali menimbulkan ketidaknyamanan, stres, dan bahkan rasa sakit terutama pada anak-anak dan pasien luka bakar. Selain itu, pemantauan RR dalam jangka waktu lama juga dapat menyebabkan iritasi dan kemerahan pada kulit. Tujuan dari penelitian ini adalah untuk mengembangkan sebuah porgram yang dapat memantau laju pernapasan secara non-kontak dan non-invasif menggunakan teknologi termografi inframerah. Hal ini didasarkan pada kenyataan bahwa terdapat perbedaan suhu pada daerah sekitar hidung pada saat inspirasi dan ekspirasi. Metode yang digunakan pada penelitian ini terdiri dari beberapa tahapan, seperti deteksi region of interest (ROI), tracking ROI, ekstraksi sinyal pernapasan, fitering dan ekstraksi nilai RR, dan klasifikasi laju pernapasan menggunakan kNN menjadi 3 kondisi yaitu takipnea, normal, dan bradipnea. Untuk memvalidasi hasilnya, deteksi RR secara kontak menggunakan photoplethysmography (PPG) dicatat secara bersamaan dan digunakan sebagai ground truth. Akurasi sistem dievaluasi dengan menghitung indeks CAND, dan diperoleh hasil rata-rata indeks CAND sebesar 94,77%. Selain itu, dihitung juga metrik kinerja sistem dan menghasilkan MAE sebesar 0,97, MSE sebesar 1,36, dan MAPE sebesar 5,23%. Evaluasi kinerja pengklasifikasi k-NN ditentukan dengan menghitung akurasi pelatihan, akurasi validasi, dan akurasi pengujian, yang masing-masing diperoleh sebesar 100%, 100%, dan 99,2%. Metrik kinerja lainnya seperti Sensitivitas, Spesifisitas, Presisi, dan pengukuran F juga dihitung untuk setiap kelas secara terpisah.
========================================================================================================
Monitoring vital physiological parameters such as respiration rate (RR) is crucial for assessing a patient’s health condition. However, current measurement techniques often require placing sensors on the patient’s body, causing discomfort, stress, and even pain, especially for children or long-term burn patients. Prolonged contact monitoring can also lead to skin irritation and redness. This study aims to develop a program that can monitor respiration rate non-contact and non-invasively using infrared thermography technology. This is based on the fact that there are temperature differences around the nose during inspiration and expiration. The method used in this study consists of several stages, including region of interest (ROI) detection and tracking, respiratory signal extraction, filtering and RR value extraction, and RR classification using kNN into three conditions: tachypnea, normal, and bradypnea. To validate the results, contact-based RR detection using photoplethysmography (PPG) was simultaneously recorded and used as the ground truth. The accuracy of the system was evaluated by calculating the CAND index, which resulted in an average of 94.77%. Furthermore, the system's performance metrics were assessed, resulting in MAE of 0.97, MSE of 1.36, and MAPE of 5.23%. The robustness and performance of the k-NN classifier is determined by computing its Training accuracy, Validation accuracy, and Testing accuracy, obtained as 100%, 100% and 99.2%, respectively. Other performance metrics such as Sensitivity, Specificity, Precision, and F-measure values are calculated as well, for each class separately.

Item Type:	Thesis (Diploma)
Uncontrolled Keywords:	Respiratory Rate, Thermal Imaging, Inspiration, Expiration, Laju Pernapasan, Pencitraan Termal, Inspirasi, Ekspirasi
Subjects:	T Technology > T Technology (General) > T58.62 Decision support systems T Technology > TA Engineering (General). Civil engineering (General) > TA1637 Image processing--Digital techniques T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK5102.9 Signal processing.
Divisions:	Faculty of Electrical Technology > Biomedical Engineering > 11410-(S1) Undergraduate Thesis
Depositing User:	Anadya Ghina Salsabila
Date Deposited:	07 Aug 2024 07:49
Last Modified:	07 Aug 2024 07:49
URI:	http://repository.its.ac.id/id/eprint/111759

Actions (login required)

View Item