End-to-end deep learning for superoscillatory subtraction imaging

TL;DR

This paper introduces an end-to-end deep learning approach that combines superoscillatory focusing and subtraction imaging, achieving sub-100-nm resolution for super-resolution optical imaging.

Contribution

It presents a unified neural network pipeline that replaces traditional multi-step processes, enhancing superoscillatory imaging resolution beyond previous limits.

Findings

Achieved deep-subwavelength imaging resolution below 100 nm.

Integrated superoscillatory focusing and subtraction imaging into a single neural network.

Eliminated manual weighting and two-step acquisition processes.

Abstract

Breaking the diffraction limit in optical imaging is crucial for resolving subwavelength details in a wide range of applications, where superoscillatory imaging and subtraction imaging are two common strategies for surpassing conventional resolution limits. We propose an end-to-end deep learning framework that integrates superoscillatory focusing and subtraction imaging into a single jointly-optimized vectorial Debye integral neural network pipeline, eliminating the traditional two-step acquisition and manual weighting process. With this end-to-end neural network, we further improve the focusing capability of the system to the sub-100-nm regime, enabling deep-subwavelength imaging resolution.