End-to-end validation of deformation-retrieval accuracy

Integrate DP-JMRNet with a complete InSAR processing chain and determine whether its improved differential-phase fidelity produces more accurate line-of-sight deformation estimates on measured datasets with reliable deformation references.

Background

DP-JMRNet is evaluated at the complex-SAR reconstruction level using differential-phase metrics, including on measured Sentinel-1 scenes, but the study does not perform complete InSAR processing or directly assess deformation-retrieval accuracy. The authors identify this as a remaining limitation because reliable phase recovery is difficult under severe noise and low temporal coherence, and because improved reconstructed differential phase may not necessarily translate into improved physical displacement estimates. Future work is explicitly proposed to connect DP-JMRNet to end-to-end InSAR processing and test the resulting line-of-sight deformation accuracy against measured datasets with reliable references.

References

Despite these results, several limitations remain. The remaining degradation under severe noise and low temporal coherence shows that reliable phase recovery is still challenging in strongly decorrelated regions. Moreover, the study focuses on differential-phase preservation at the complex-SAR reconstruction level and does not directly evaluate deformation retrieval using a complete InSAR processing chain. Future work will therefore integrate DP-JMRNet with InSAR processing and assess whether the improved differential-phase fidelity translates into more accurate line-of-sight deformation estimates on measured datasets with reliable deformation references.

— DP-JMRNet: A Deep Unfolding Network for Differential Phase Preservation in Sparse Bitemporal SAR Reconstruction  (2608.26605 - Bao et al., 27 Aug 2026) in Section Conclusions