Neural network improves 3D imaging by accurately finding hidden objects
Multi-Level-Set-Based Physics-Driven Neural Network to Solve 3-D Inverse Scattering Problems
Machine Learning
Summary
Finding objects inside a 3D space using electromagnetic waves is hard because images can get blurry and unclear. The authors designed a special kind of neural network that uses multiple layers to represent different materials and shapes more clearly. Their method helps reconstruct clearer boundaries and more uniform materials while reducing noise and errors in the background. This approach works well even when objects have complicated shapes or are close together, and it adapts to varying measurement conditions. Overall, their technique enhances 3D imaging by improving how objects are detected and shown.
Inverse scatteringNeural networkLevel set methodElectromagnetic imaging3D reconstructionTotal variation regularizationMaterial contrastSoft-union modelPhysics-driven optimizationComplex permittivity
Authors
Yutong Du, Zicheng Liu, Bo Qi, Yali Zong, Peixian Han
Abstract
This paper proposes a level-set-based physics-driven neural network solver (LSPDNN) for 3-D electromagnetic inverse scattering. To mitigate boundary blurring and reconstruction artifacts in voxel-wise contrast reconstruction, the proposed solver exploits the piecewise homogeneity of practical scatterers by representing unknown targets with multiple coordinate-dependent neural level-set components. Specifically, a soft-union multi-material model is proposed to separately describe the object support and material distribution. The global support is formed by the union of multiple level-set components, while the local contrast is determined by normalized component weights and learnable complex permittivity candidates. In addition, a model-consistent total variation (TV) regularization is imposed on the material-region indicators, rather than directly on the reconstructed contrast, to suppress fragmented material assignments without excessively smoothing material interfaces. An adaptive loss balancing strategy is further introduced to reduce the dependence on manually selected regularization weights. For each measurement instance, the neural level-set parameters and material candidates are optimized by minimizing a physics-consistent objective function. Numerical and experimental results demonstrate that LSPDNN can reconstruct scatterers with clear boundaries, more uniform material regions, and substantially reduced background artifacts. The results highlight the advantage of the neural level-set parameterization in challenging 3-D inverse scattering cases involving irregular shapes, closely spaced objects, multiple materials, and measurement noise.