Creating a Virtual Population of the Human Nasal Cavity forVelocity-Based Predictions of Respiratory Flow Features Using GraphConvolutional Neural Networks

GNNs can be applied to any shape or volume represented by a graph, e.g., triangulated shapes, or computational grids. Convolutional filters in GNNs operate on nodes and their neighboring nodes. This allows more efficient training compared to convolutional neural networks (CNNs), whose convolutional filters are rectangular and operate in Cartesian directions. The goal is to predict respiratory system flow features such as air resistance, wall shear stress, and energy flux within the human nasal cavity during inspiration. The initial step involves generating a virtual population through random scaling applied simultaneously to length, width, and height. Three distinct geometries are chosen to generate 297 virtual patients, including an average one based on 35 healthy patients, a Caucasian healthy patient, and an Asian healthy patient. The second part of the talk exposes the preliminary results based on 297 patients with physiological observations and discussions on the accuracy result of the GCNN model.