Cross-Dataset Generalization in Deep Learning

TL;DR

This paper investigates the generalization challenge in deep learning for imaging through scattering media, revealing that dataset diversity enhances model robustness across different data sources.

Contribution

It demonstrates that increasing training dataset diversity improves the approximation of true physical mappings, aiding cross-dataset generalization in deep learning models.

Findings

Enhanced dataset diversity improves generalization.

The learned relationship is dataset-dependent, not the true physical model.

Insights into designing training datasets for better generalization.

Abstract

Deep learning has been extensively used in various fields, such as phase imaging, 3D imaging reconstruction, phase unwrapping, and laser speckle reduction, particularly for complex problems that lack analytic models. Its data-driven nature allows for implicit construction of mathematical relationships within the network through training with abundant data. However, a critical challenge in practical applications is the generalization issue, where a network trained on one dataset struggles to recognize an unknown target from a different dataset. In this study, we investigate imaging through scattering media and discover that the mathematical relationship learned by the network is an approximation dependent on the training dataset, rather than the true mapping relationship of the model. We demonstrate that enhancing the diversity of the training dataset can improve this approximation, thereby achieving generalization across different datasets, as the mapping relationship of a linear physical model is independent of inputs. This study elucidates the nature of generalization across different datasets and provides insights into the design of training datasets to ultimately address the generalization issue in various deep learning-based applications.