Implementation of Image Quality Improvement Methods and Lung Segmentation on Chest X-Ray Images Using U-Net Architectural Modifications

COVID-19 is an infectious disease that causes acute respiratory distress syndrome due to the SARS-CoV-2 virus. Rapid and accurate screening and early diagnosis of patients play an essential role in controlling outbreaks and reducing the spread of this disease. This disease can be diagnosed by manually reading CXR images, but it is time-consuming and prone to errors. For this reason, this research proposes an automatic medical image segmentation system using a combination of U-Net architecture with Batch Normalization to obtain more accurate and fast results. The method used in this study consists of pre-processing using the CLAHE method and morphology opening, CXR image segmentation using a combination of U-Net-4 Convolution Block architecture with Batch Normalization, then evaluated using performance measures such as accuracy, sensitivity, specificity, F1-score, and IoU. The results showed that the U-Net architecture modified with Batch Normalization had successfully segmented CXR images, as seen from all performance measurement values above 94%.

Classification of Covid-19 Diseases Through Lung CT-Scan Image Using the ResNet-50 Architecture

Covid-19 is a disease of the respiratory tract caused by the Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) virus. One way to diagnose Covid-19 can be done by examining lung abnormalities on the results of a Computed Tomography Scan (CT-Scan) of the lungs. However, the determination of the diagnostic results obtained requires high accuracy and quite a long time. For this reason, an automatic system is needed to make it easier for medical personnel to diagnose Covid-19 disease quickly and accurately. One way to do this with the help of a computer is pattern recognition. In this study, pattern recognition techniques were used which were divided into three stages, namely pre-processing, feature extraction and classification. The methods used in the pre-processing stage are grayscale and Contrast Limited Adaptive Histogram Equalization (CLAHE) to improve image quality and contrast. The extraction stage uses the Principal Component Analysis (PCA) method, because it can reduce data dimensions without eliminating important features in the data. For the classification stage, a deep learning-based method is used, namely the Convolutional Neural Network (CNN). The CNN architecture used in this study is Resnet-50. The method proposed in this research is evaluated by measuring the performance values of accuracy, recall, precision, F1-score, and Cohen Kappa. The results of the study indicate that the PCA method has worked optimally in dimension reduction, without losing important features on CT-scan images of the lungs. Besides that, the proposed method has succeeded in classifying Covid-19 very well, as seen from the accuracy, Recall, Precision, F1-Score and Cohen Kappa values above 90%.