Discontinuous Galerkin Neural Operator for Pathology Defocus Deblurring

Researchers introduced DGNO, a Discontinuous Galerkin Neural Operator for pathology microscopy deblurring.

The model treats optical blur as an integral operator with local elements and interface fluxes, making it well suited to spatially varying defocus.

DGNO reportedly outperforms existing methods on heterogeneous defocus blur, producing sharper restorations.

The approach targets a difficult medical imaging problem and may offer more interpretable reconstructions than standard deep learning models.