A comprehensive survey on chest diseases analysis: technique, challenges and future research directions

Learning with chest diseases and their classification, segmentation, localization, annotation, and abnormality detection are challenging and exciting research objectives. Over the past few years, different researchers have come with various learning techniques for improving performance in chest image analysis. However, the scarcity of labeled datasets and less computational processing power was a reason for the negligible performance improvement. Nevertheless, with the advancement in Deep Learning (DL) techniques, researchers succeeded in achieving state-of-the-art results and created a new research paradigm. Among the different DL techniques, Convolutional Neural Network comes with a revolution for identifying abnormality in chest images. This survey aims to highlight the importance of deep learning techniques in chest disease diagnosis. In this paper, our primary objective is to broadly analyze different DL techniques and recognize some of the important research challenges that mostly affect deep neural networks for investigating various chest diseases. Specifically, we focus on several chest diseases, symptoms, preliminary treatments, and state-of-the-art detection techniques. We also introduce several chest image analysis tools, techniques, and datasets for analyzing chest diseases. Further, we have presented several open research challenges and future research directions in the field of chest image analysis.