Efficient sparse-view medical image classification for low radiation and rapid COVID-19 diagnosis

TL;DR

A new deep learning model called ProMAE can diagnose COVID-19 from sparse CT scans with high accuracy, reducing radiation exposure and enabling faster diagnosis.

Contribution

ProMAE introduces a column-wise masking strategy for learning diagnostic features directly from sparse-view sinograms without reconstruction.

Findings

ProMAE achieves over 95% diagnostic accuracy at all sparsity levels up to 99%.

ProMAE outperforms ResNet, ConvNeXt, and conventional MAE models in high sparsity environments.

The model enables accurate diagnosis with minimal radiation exposure, suitable for portable imaging systems.

Abstract

This study proposes a deep learning-based diagnostic model called the Projection-wise Masked Autoencoder (ProMAE) for rapid and accurate COVID-19 diagnosis using sparse-view CT images. ProMAE employs a column-wise masking strategy during pre-training to effectively learn critical diagnostic features from sinograms, even under extremely sparse conditions. The trained ProMAE can directly classify sparse-view sinograms without requiring CT image reconstruction. Experiments on sparse-view data with 50%, 75%, 85%, 95%, and 99% sparsity show that ProMAE achieves a diagnostic accuracy of over 95% at all sparsity levels and, in particular, outperforms ResNet, ConvNeXt, and conventional MAE models in COVID-19 diagnosis in environments with 85% or higher sparsity. This capability is especially advantageous for the development of portable and flexible imaging systems during large-scale outbreaks such as COVID-19, as it ensures accurate diagnosis while minimizing radiation exposure, making it a vital tool in resource-limited and high-demand settings.

Figures