Identifikasi Gangguan Pernapasan Menggunakan Sensor Thermal Dengan Metode Convolutional Neural Network (CNN)

Erianthy, Cicilia Selvy (2021) Identifikasi Gangguan Pernapasan Menggunakan Sensor Thermal Dengan Metode Convolutional Neural Network (CNN). Undergraduate thesis, Politeknik Negeri Jember.

	Text (Abstrak) Abstract.pdf - Accepted Version Available under License Creative Commons Attribution Non-commercial Share Alike. Download (7kB)
	Text (Bab 1) Bab 1.pdf - Accepted Version Available under License Creative Commons Attribution Non-commercial Share Alike. Download (12kB)
	Text (Daftar Pustaka) Daftar Pustaka.pdf - Accepted Version Available under License Creative Commons Attribution Non-commercial Share Alike. Download (235kB)
	Text (Laporan Lengkap) Laporan lengkap.pdf - Accepted Version Restricted to Registered users only Download (0B) | Request a copy

Abstract

In 2020, Indonesia was affected by the Covid-19 disease outbreak. Covid-19 is a disease caused by the coronavirus which is currently a pandemic occurring in many countries. Recent studies have found that Covid-19 patients have obvious respiratory symptoms such as shortness of breath, fever, fatigue, and dry cough (Sohrabi et al., 2020; Xu et al., 2020). Among these symptoms, irregular breathing is considered as one of the early signs of Covid-19 (Chow et al., 2020). For many people, the initial symptoms of mild breathing are difficult to identify. Conventional respiration measurements usually place the sensor by placing it on the chest (AlKhalidi et al., 2011). Contact measuring equipment is large, expensive, and the measurement process takes a long time. Also, contact with suspected Covid-19 during measurement can increase the risk of spreading Covid-19. Therefore, noncontact measurement is more suitable for the current situation. In recent years, many non-contact respiration measurement methods have been developed based on image sensors, doppler radar (Kranjec et al., 2014), depth cameras (Wang et al., 2020) and thermal cameras (Hu et al., 2017). . Based on these problems, the researchers classified the features of the thermal signal using the Convolution Neural Network (CNN) method to determine the presence or absence of respiratory disorders. The results of the study using an image input of 60 x 60 pixel, the learning rate value of 0.001, as many as 10 epochs with training and testing data of 24 and 8 data. Resulting in an accuracy rate of training model reaching 95 percent and testing reaching 87 percent

Actions (login required)

View Item