Search papers, labs, and topics across Lattice.
This paper introduces a novel deep learning framework for restoring images affected by multiplicative Gamma noise and blur, utilizing an explicit regularizer informed by geometric priors such as surface area and mean curvature. The authors develop a mirror descent algorithm optimized for Gamma-noise fidelity terms, ensuring global convergence to critical points through the Kurdyka-Lojasiewicz property. Experimental evaluations reveal that this approach not only surpasses traditional model-based methods but also matches the performance of leading deep equilibrium models while significantly reducing the number of trainable parameters.
Explicit geometric priors in image restoration can outperform traditional methods while using fewer parameters, challenging the reliance on implicit regularization.
We propose a deep learning framework for image restoration from images degraded by both multiplicative Gamma noise and blur. Unlike conventional deep equilibrium (DEQ) models that rely on implicit neural regularization, the proposed method learns an explicit and interpretable regularizer parameterized by geometric priors associated with surface area and mean curvature. To minimize the resulting variational model, we develop a mirror descent algorithm tailored to the commonly used Gamma-noise fidelity terms. Leveraging the Kurdyka-Lojasiewicz property for functions defined in $o$-minimal structures, we establish the global convergence of the generated iterates to a critical point. Experimental results on both grayscale and color image restoration demonstrate that the proposed method consistently outperforms representative model-based approaches while achieving performance comparable to state-of-the-art DEQ models based on implicit regularization, despite requiring substantially fewer trainable parameters.
