Segmentation and classification on chest radiography: a systematic survey

Abstract

Chest radiography (X-ray) is the most common diagnostic method for pulmonary disorders. A trained radiologist is required for interpreting the radiographs. But sometimes, even experienced radiologists can misinterpret the findings. This leads to the need for computer-aided detection diagnosis. For decades, researchers were automatically detecting pulmonary disorders using the traditional computer vision (CV) methods. Now the availability of large annotated datasets and computing hardware has made it possible for deep learning to dominate the area. It is now the modus operandi for feature extraction, segmentation, detection, and classification tasks in medical imaging analysis. This paper focuses on the research conducted using chest X-rays for the lung segmentation and detection/classification of pulmonary disorders on publicly available datasets. The studies performed using the Generative Adversarial Network (GAN) models for segmentation and classification on chest X-rays are also included in this study. GAN has gained the interest of the CV community as it can help with medical data scarcity. In this study, we have also included the research conducted before the popularity of deep learning models to have a clear picture of the field. Many surveys have been published, but none of them is dedicated to chest X-rays. This study will help the readers to know about the existing techniques, approaches, and their significance.

Revisiting Chest X-ray potency as a gold standard candidate for early screening of clinical diagnosis for COVID-19 infection

“…However, the diagnostic performance of chest X-ray in the early stages of disease is very limited, as some pathological findings on chest computed tomography (CT) may not be detectable on X-ray. [30][31][32] The findings of this study confirm the poor performance of the diagnostic chest Xray if it is based on the conclusion that pneumonia is present or not. Based on this, it is still necessary to develop a more detailed study regarding the comparison of the pattern of alveolar damage due to COVID-19 infection in cases with mild, moderate and severe degrees, so that these differences can be identified as diagnostic target outcomes.…”

Computer-assisted diagnosis for an early identification of lung cancer in chest X rays

“…For the preparation of the CAD algorithm used in the present study, a hierarchical classifier addressed the characteristics of the image in sequential order, first discerning posteroanterior chest images from other types of x-rays, to subsequently identify which of them are at high probability for pulmonary nodule, and in these stages the attained accuracies were 96% and 79%, respectively. Previous reports on CAD algorithms used for the identification of nodules have shown similar results 29 , 31 – 33 . The usefulness of the algorithm used in the present study for the identification of pulmonary nodules was confirmed by the finding that high probability images according to the algorithm had a nodule confirmed by the radiologist in 23.2% of the cases, while images with lower probabilities only show nodules in less than ten percent of the cases, a difference that was statistically significant.…”

“…Computer aided diagnosis (CAD) algorithms may be integrated in diagnostic software and have been successfully used in thoracic imaging 24 – 26 , including chest x-ray assessment 27 – 30 , with accuracies over 90% for the identification of pulmonary nodules in some studies 31 , 32 . However, databases used until now for the creation of these algorithms have been in most cases small and with limited capabilities of machine learning 31 , 33 – 36 . Furthermore, studies that have been focused on the identification of pulmonary nodules in chest x-rays through CAD algorithms have not included a clinical validation of the radiologic diagnoses, an important point considering that part of these nodules would be malignant, and their early management before the appearance of symptoms would improve their prognosis.…”

OWAE-Net: COVID-19 detection from ECG images using deep learning and optimized weighted average ensemble technique

“…Therefore, the researchers use the transfer learning technique to train their models. Different research groups have released pre-trained models for reuse Agrawal and Choudhary (2022) . Most of these pre-trained models available are trained on the ImageNet dataset Deng et al.…”