UniPolar: Joint Image and Polarimetric Restoration under Compound Degradations via Physical Consistency Learning
To address this challenge, we propose UniPolar, a coupled Image Stokes framework for joint image polarimetric restoration from compound-degraded CPFA observations.
ProofPaper ↗
Key points
- Polarimetric image restoration aims to recover polarization angle images and reliable polarimetric quantities from degraded observations.
- Beyond visual fidelity, it also requires preserving the physical relationships underlying reliable polarimetric estimation.
- Such compound degradations jointly perturb image intensities and cross angle polarization differences.
- These results underscore the importance of jointly considering visual fidelity and polarimetric reliability, offering a physically grounded perspective on learning based restoration under compound degradations.
Sources (1)
- [1]UniPolar: Joint Image and Polarimetric Restoration under Compound Degradations via Physical Consistency LearningarXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 16, 12:07 PM
To address this challenge, we propose UniPolar, a coupled Image Stokes framework for joint image polarimetric restoration from compound-degraded CPFA observations.
Polarimetric image restoration aims to recover polarization angle images and reliable polarimetric quantities from degraded observations.
Extractive summary: sentences quoted from the sources.