Interpretable Classification of Pneumonia Infection Using Explainable AI (XAI-ICP)

In contrast, the system was tested and trained at Taichung Veterans General Hospital (TCVGH) because many medical examiners find open-box models in the medical field to be highly desirable. The thorough debate regarding the outcome is semi-transparent in the working process, so even though the accuracy predicted by most convolutional neural network (CNN) models is good, it is still not convincing. Because most people have weakened immune systems, pneumonia is one of the most common and highly contagious illnesses. Therefore, this paper’s objective is to use explainable AI (XAI-ICP) to create an interpretable categorization of pneumonia infection. Therefore, the extremely effective XAI- ICP system was created to address this problem by adjusting to the current state of population health. The aim is to design an interpretable deep classification and transfer learning-based evaluation for pneumonia infectionclassification. The model is primarily pre-trained using the open Chest X-Ray (CXR) dataset from NationalInstitutes of Health (NIH). Independent learning, Taiwan VinDr open dataset for transfer learning ofpneumonia affected patients with labeled CXR images possessing three features of infiltrate, cardiomegalyand effusion. The data labeling is performed by the medical examiners with the XAI human-in-the-loop approach. XAI-ICP demonstrates the XAI based reconfigurable DCNN with human-in-the- loop as a novelapproach.