Deep learning (DL)-based holographic microscopy has emerged as a powerful label-free imaging modality that enables capturing intricate tissue details at the cellular level. While DL methods have shown considerable advances in phase retrieval performance over classical methods, they face notable challenges when dealing with out-of-distribution (OOD) data, which arise from variations in tissue types and imaging system parameters. These challenges limit the practical applicability of DL-based phase retrieval in real-world scenarios.

The cover image for Advanced Photonics Nexus Volume 4 Issue 2 visually represents a proposed holographic reconstruction process where complex phase information is transferred from tissue samples to the sensor and then reconstructed. This method is specifically designed to adapt to all variables inherent in imaging systems and tissue characteristics, ensuring reliable phase retrieval under highly perturbed conditions. The approach marks a significant step toward a generalized DL-based phase retrieval scheme in histopathology. For details, see the Gold Open Access article by Jiseong Barg, Chanseok Lee, Chunghyeong Lee, and Mooseok Jang, “Adaptable deep learning for holographic microscopy: a case study on tissue-type and system variability in label-free histopathology,” Adv. Photon. Nexus 4(2) 026005 (2025), DOI: 10.1117/1.APN.4.2.026005