aiqbench analysis
This research introduces FoundCAC, a foundational framework that addresses generalization limits in blind lens aberration correction by leveraging a large-scale, unbiased lens library and discrete degradation priors. By encoding continuous Point Spread Functions into a latent representation, the approach achieves state-of-the-art zero-shot performance while enabling efficient few-shot adaptation for unseen optical systems. The methodology offers a scalable solution for handling diverse unknown degradations, with code and datasets slated for public release.
Towards Blind Lens Aberration Correction via Large LensLib Pre-training and Discrete Degradation Priors
arXiv:2511.17126v4 Announce Type: replace-cross Abstract: Emerging deep-learning-based lens library pre-training (LensLib-PT) pipeline offers a new avenue for blind lens aberration correction by training a universal neural network, demonstrating strong capability in handling diverse unknown optical degradations. This work proposes FoundCAC, a universal foundational framework that resolves two challenges hindering the generalization of existing pipelines: the difficulty of scaling training data and the absence of prior guidance characterizing optical degradation. To improve data scalability, we expand the design specifications to increase degradation diversity and construct AODLibpro, a large-scale, unbiased lens library based on a uniform sampling strategy that quantifies spatial-variation patterns and severity. In terms of model design, to leverage Point Spread Functions (PSFs) as guidance while maintaining the blind paradigm, we propose a multi-stage vector-quantized representation learning scheme. This paradigm is specifically designed to construct a Latent PSF Representation (LPR), explicitly encoding complex continuous PSFs into a discrete degradation prior to regularize the highly ill-posed restoration process. Through a simple yet effective codebook-freezing strategy, our framework leverages the discrete prior to elevate full-shot restoration performance and unlock highly efficient few-shot adaptation for unseen lenses. Experiments on diverse aberrations of synthetic LensLib and real-world lenses demonstrate that our framework achieves state-of-the-art zero-shot generalization while enabling highly efficient few-shot adaptation for specific lenses. The source code and datasets will be made publicly available at https://github.com/zju-jiangqi/FoundCAC.
Abstract:Emerging deep-learning-based lens library pre-training (LensLib-PT) pipeline offers a new avenue for blind lens aberration correction by training a universal neural network, demonstrating strong capability in handling diverse unknown optical degradations. This work proposes FoundCAC, a universal foundational framework that resolves two challenges hindering the generalization of existing pipelines: the difficulty of scaling training data and the absence of prior guidance characterizing optical degradation. To improve data scalability, we expand the design specifications to increase degradation diversity and construct AODLibpro, a large-scale, unbiased lens library based on a uniform sampling strategy that quantifies spatial-variation patterns and severity. In terms of model design, to leverage Point Spread Functions (PSFs) as guidance while maintaining the blind paradigm, we propose a multi-stage vector-quantized representation learning scheme. This paradigm is specifically designed to construct a Latent PSF Representation (LPR), explicitly encoding complex continuous PSFs into a discrete degradation prior to regularize the highly ill-posed restoration process. Through a simple yet effective codebook-freezing strategy, our framework leverages the discrete prior to elevate full-shot restoration performance and unlock highly efficient few-shot adaptation for unseen lenses. Experiments on diverse aberrations of synthetic LensLib and real-world lenses demonstrate that our framework achieves state-of-the-art zero-shot generalization while enabling highly efficient few-shot adaptation for specific lenses. The source code and datasets will be made publicly available at this https URL.
Accepted to 2026 IEEE International Conference on Computational Photography (ICCP). The source code and datasets will be made publicly available at this https URL
Image and Video Processing (eess.IV); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG); Optics (physics.optics)
arXiv:2511.17126 [eess.IV]
(or arXiv:2511.17126v4 [eess.IV] for this version)
Full text under the publisher's open licence · 268 words
Related coverage
Game Design Creativity evals
LessWrong · Oct 12, 2026
Does DiffusionGemma do latent reasoning?
AI Alignment Forum · Aug 16, 2026
The J-Space Debate, Agent Swarms, and Pacing Frontier AI - Digital Minds Newsletter #4
LessWrong · Sep 18, 2026
News Archive 2026
Quantum Computing Report · Jul 25, 2026