| |
|
| |
Aurthors/المؤلفون
Abstract/الملخص
Keywords/الكلمات المفتاحية
Content/أقسام الملف
Introduction
Materials and methods
Aim and Objectives
References |
| The Use of Image Processing Techniques in Diagnosing Chest Disease |
| |
| |
| Dr. Mohammad Issa(1) Dr. Ghada Saad(2) Aous Mohammad*(3) Noman Dali (4) Arwa Ahmad (5) |
1. Phd: Biomedical Department - Faculty Biomedical Engineering-Al-Andalus University for Medical Science–Syria.
2. Phd: Biomedical Department - Faculty Biomedical Engineering-Al-Andalus University for Medical Science–Syria.
3. Master's in the Department of Computer and Automatic control Engineering of faculty of Mechanical and electrical Engineering - Tishreen University.
4. Biomedical Engineering, Faculty Biomedical Engineering- -Al-Andalus University for Medical Science –Syria
5. Biomedical Engineering, Faculty Biomedical Engineering- -Al-Andalus University for Medical Science –Syria |
| Email of the corresponding: aosmohamed93@gmail.com |
| |
| |
|
|
Abstract
Chest diseases are considered one of the most dangerous chronic diseases in the world. The delay in their diagnosis and the initiation of treatment is one of the most important causes of death nowadays. The aim of this paper was to develop a working methodology for diagnosing chest diseases automatically using an algorithm based on processing CT scan images of the chest for people with COVID-19. This methodology included the following approaches: datasets from kaggle, initial processing of CT images, segmenting the images using different image processing techniques, moving to the testing phase of many neural networks until the inception-v3 network has been selected. Ultimately, training this network to reach an accuracy of 71.41% in order to obtain an accurate diagnosis of the patient’s condition whether sick or healthy as COVID-19 pandemic is a major hotspot for current research and clinical studies. |
|
|
|
| Keywords: Chest diseases, CT images, COVID-19, neural networks, kaggle datasets. |
|
| |
|
| |
|
| Introduction |
|
Nowadays, the need for speed and accuracy in the diagnosis of chest diseases has increased rapidly worldwide. COVID-19 pandemic is currently diagnosed through the Polymerase Chain Reaction (PCR) technique. However, it may take up to two days to diagnose the disease. The test is, sometimes, repeated to rule out any possibility of errors’ results [1]. Hence, this paper aims to speed up the diagnosis of lung diseases and the degree to which the disease has reached, automatically and quickly, by relying on image processing techniques to reach an accurate diagnosis. Coronavirus disease 2019 has been named (COVID-19), identified in Wuhan, China, as (CO) stands for corona, (VI) for virus and ‘D’ for disease. It was referred to as ‘2019 novel coronavirus’ or ‘2019-nCoV’ [2]. Among these diagnostic techniques are X-Ray Imaging and Magnetic Resonance Imaging (MRI) and Computed Tomography (CT) scan. The most important of which is CT scan so as most of the signs of COVID-19 are visible in the CT images [3]. These appear as white spots, indicating ‘ground-glass opacities’ which become more intense over time; similar to those found in SARS or Middle East Respiratory Syndrome [4]. Recent studies have shown the rapid development in the use of image processing techniques and artificial intelligence in the development of CAD systems for automated diagnosis of COVID-19 detection [5]. Hence, the importance of this paper is due to the fact that it provides algorithms that help doctors and physicists to predict and diagnose chest diseases after the global widespread of COVID-19 pandemic.
Recently, various techniques have been developed by researchers. Some examples of these are Otsu's technique to extract the region of interest, Wavelet transform, histogram normalization such as image filtering techniques, feature extraction with GLCM algorithm, image composition analysis, and classification of fuzzy logic neural networks SVM [6] [7] [8] [9] [10]. There is a common similarity in the literature in dealing with different methods of image processing for the diagnosis of some chest diseases. However, with the emergence of COVID-19 pandemic infection, there is, therefore, an urgent need to conduct and undertake more research and clinical studies to reach algorithms having the capabilities of aiding in the diagnosis of these diseases automatically and quickly in a precise manner. |
|
| |
|
| Materials and methods |
|
| The suggested methodology has been for extracting lung region using CT images based on Otsu’s single-thresholding used to perform automatic image thresholding. In its simplest form, the algorithm returns a single intensity boundary that separates image elements into two classes, foreground and background. This threshold is determined by minimizing the intensity variance within the category, or equivalently, by maximizing the inter-category variance [9] [10] [11]. The main characteristics that distinguish the COVID-19 disease were revealed, and therefore, a distinction has been made between healthy images and images with the disease using a CNN classifier. |
|
| Aim and Objectives |
|
The aim of this paper has been to develop a working methodology for diagnosing chest diseases automatically and quickly using an algorithm based on processing CT scan images of the chest for people infected with COVID-19. The main objectives can be summarised as follows:
1- Fragmentation of CT images to extract the region of the lung among other tissues;
2- Classification of diseased tissues within the lung region among the healthy tissues;
3- If the diseased areas are exposed, they are classified as injured, otherwise they are healthy.
The methodology has been developed in a box scheme, based on securing the database, initialling and processing the CT images and fragmenting these images to be entered into a neural network to train this network with the aim of diagnosing COVID-19 infection. Figure (1) shows the box diagram of the algorithm used |
|
| |
|
 |
|
| Figure 1 The box diagram of the algorithm |
|
It has been worked on an ASUS computer with the following specifications:
Processor: Intel® Core™ i7-8750H CPU @ 2.20GHz (12 CPUs), ~2.2Ghz
Memory: 16384MB RAM |
|
| References |
|
1. Ammar M. (2013). Medical Image Processing and Display Systems. Damascus University Publications. Damascus, Syria.
2. Pollard, C. A., Morran, M. P., & Nestor-Kalinoski, A. L. (2020). The COVID-19 pandemic: a global health crisis. Physiological genomics, 52(11), 549-557.
3. Awulachew, E., Diriba, K., Anja, A., Getu, E., & Belayneh, F. (2020). Computed tomography (CT) imaging features of patients with COVID-19: systematic review and meta-analysis. Radiology Research and Practice, 2020.
4. Cereser, L., Da Re, J., Zuiani, C., & Girometti, R. (2021). Chest high-resolution computed tomography is associated to short-time progression to severe disease in patients with COVID-19 pneumonia. Clinical Imaging, 70, 61-66.
5. Abbasian Ardakani, A., Acharya, U. R., Habibollahi, S., & Mohammadi, A. (2021). COVIDiag: a clinical CAD system to diagnose COVID-19 pneumonia based on CT findings. European radiology, 31(1), 121-130.
6. Khambampati, A. K., Liu, D., Konki, S. K., & Kim, K. Y. (2018). An automatic detection of the ROI using Otsu thresholding in nonlinear difference EIT imaging. IEEE Sensors Journal, 18(12), 5133-5142.
7. Shalchian, B., Rajabi, H., & Soltanian-zadeh, H. (2009). Fusion of PET and CT images using wavelet transform. Hellenic journal of nuclear medicine, 12(3), 238-243.
8. Pisano, E. D., Zong, S., Hemminger, B. M., DeLuca, M., Johnston, R. E., Muller, K., ... & Pizer, S. M. (1998). Contrast limited adaptive histogram equalization image processing to improve the detection of simulated spiculations in dense mammograms. Journal of Digital imaging, 11(4), 193-200.
9. Mall, P. K., Singh, P. K., & Yadav, D. (2019, December). GLCM based feature extraction and medical X-RAY image classification using machine learning techniques. In 2019 IEEE Conference on Information and Communication Technology (pp. 1-6). IEEE.
10. Huang, S., Cai, N., Pacheco, P. P., Narrandes, S., Wang, Y., & Xu, W. (2018). Applications of support vector machine (SVM) learning in cancer genomics. Cancer genomics & proteomics, 15(1), 41-51.
11. Aous M., Ghada S. Designing a Multiclassification Convolutional Neural Networks Model for the Diagnosis of Lung Cancer and Covid-19. Tishreen University Journal for Research and Scientific Studies - Engineering Sciences Series. 2022; 44(6):185-20. |
|
| |
|
| |
|
| |
|
|
