Flexible Granular Deep Learning Voting Architecture

The recognition of sleep stages is a very complex and novel topic for many physicians. In an attempt to capitalise on the advances in this field of computing, a number of machine learning algorithms have attempted to replicate the actions that physicians conduct. Deep learning researchers are attempting to create and assess a variety of models and building blocks to assist physicians with their analysis because the polysomnograph's structure is a challenging combination of reasonably comparable components. In this work, we present a broad yet adaptable architecture that leverages the main driving force behind voting systems to enable scientists and doctors to incorporate as many criteria or blocks as required.