DeepFilters: Scattering-Aware Pupil Engineering with Learned Digital Filter Reconstruction for Extended Depth of Field Microscopy

2026-05-13

AI Summary

This work addresses the severe performance degradation of conventional extended-depth-of-field microscopy in scattering biological tissues, where deep-tissue signal recovery remains challenging. The authors propose DeepFilters, a novel framework that, for the first time, integrates scattering-aware mechanisms into deep optical design by jointly optimizing a parametric pupil filter and a digital reconstruction network. Leveraging a differentiable imaging model, the approach achieves robustness to scattering and generalizes across diverse samples without retraining. By combining physics-guided regularization with a hybrid genetic–gradient optimization strategy, the method extends the point spread function beyond 400 µm in transparent media and enables effective imaging at depths exceeding 120 µm in biological tissue. Validation on fixed brain slices and sea urchin embryos demonstrates significant improvements in both imaging depth and stability.

Abstract

Extended depth of field microscopy encodes axial information into a single acquisition through engineered point spread functions, but conventional and deep optics approaches are subject to degradation in scattering tissue. We introduce DeepFilters, a scattering-aware deep optics framework that jointly optimizes a parameterized pupil filter and a digital-filter-based reconstruction network through a calibrated differentiable forward model to achieve broad generalization without retraining. Incorporating empirical scattering kernels, physics-guided regularization, and a hybrid genetic-gradient initialization strategy, DeepFilters extends the PSF from 16 micron to >400 micron in clear media and enables signal recovery beyond 120 micron deep in biological tissues, validated across fixed brain slices and sea urchin embryos.

Problem

• extended depth of field microscopy
• scattering tissue
• point spread function degradation
• deep imaging
• signal recovery

Innovation

• scattering-aware optics
• pupil engineering
• digital filter reconstruction
• extended depth of field
• differentiable forward model