A Neoteric Variant of Deep Learning Network for Chest Radiograph Automated Annotation

Abstract

Automated annotation and classification of chest radiographs is the pressing need for modern biomedical technologies. This is mainly because of the massive volume of radiograph archives. The variants of machine learning models have handled this issue of automated disease annotation. However, the performance is found to be constrained due to the visual attribute dependency. Here, deep learning has come into the focus to submit the contribution for effective and efficient automated disease annotation. In this paper, a new variant of a deep learning network (DLN) is presented for automated annotation. Moreover, the exhaustive parametric comparison of the variant with the classical network and the pre-trained network is presented. The Chest X pert dataset is considered for this comparative study. The simulation results advocated for the effectiveness of devised variants.