Physics-informed Shadowgraph Network: An End-to-end Density Field Reconstruction Method

This addresses the need for accurate density field reconstruction in fluid dynamics or imaging applications, representing an incremental advancement by integrating physics constraints into neural networks.

The paper tackles the problem of reconstructing density fields from shadowgraph images by introducing a physics-informed neural network method, achieving quantitative reconstruction as reported.