Pulmonary Embolism Detection Using Machine and Deep Learning Techniques

Summary

Pulmonary embolism (PE) is a disease that occurs due to blood clot in pulmonary arteries. Various methods are available for pulmonary embolism detection. The primary method to detect PE is computed tomography pulmonary angiogram (CTPA). One CTPA contains hundreds of images. It is very difficult for a radiologist to study such a CTPA on time. The radiologist's accuracy and efficiency are also affected by other human factors such as attention span and eye fatigue. This delayed study of PE detection may cause serious illness or death. Many research efforts have been made to solve this problem with computer‐aided detection (CAD) systems. The pulmonary embolism CAD system helps the radiologist for a better and timely diagnosis. This system uses light from various machines and deep learning‐based CAD models designed for PE detection. In PE detection, we can perform classification, detection, and segmentation. A comparative analysis of various proposed models of PE detection is also conducted in this study.