OmniLens: Towards Universal Lens Aberration Correction via LensLib-to-Specific Domain Adaptation

TL;DR

OmniLens introduces a universal lens aberration correction framework that leverages a comprehensive lens library and domain adaptation to improve image quality across diverse lenses, including unseen ones.

Contribution

It proposes a novel LensLib-to-specific domain adaptation method with an EAOD pipeline for generating diverse lens samples, enhancing generalization in aberration correction.

Findings

Universal CAC model outperforms existing methods in PSNR by 0.35-1.81dB.

Domain adaptation significantly improves performance, especially in severe aberration cases.

EAOD-generated LensLib effectively supports robust model training.

Abstract

Emerging universal Computational Aberration Correction (CAC) paradigms provide an inspiring solution to light-weight and high-quality imaging with a universal model trained on a lens library (LensLib) to address arbitrary lens optical aberrations blindly. However, the limited coverage of existing LensLibs leads to poor generalization of the trained models to unseen lenses, whose fine-tuning pipeline is also confined to the lens-descriptions-known case. In this work, we introduce OmniLens, a flexible solution to universal CAC via (i) establishing a convincing LensLib with comprehensive coverage for pre-training a robust base model, and (ii) adapting the model to any specific lens designs with unknown lens descriptions via fast LensLib-to-specific domain adaptation. To achieve these, an Evolution-based Automatic Optical Design (EAOD) pipeline is proposed to generate a rich variety of lens samples with realistic aberration behaviors. Then, we design an unsupervised regularization term for efficient domain adaptation on a few easily accessible real-captured images based on the statistical observation of dark channel priors in degradation induced by lens aberrations. Extensive experiments demonstrate that the LensLib generated by EAOD effectively develops a universal CAC model with strong generalization capabilities, which can also improve the non-blind lens-specific methods by 0.35~1.81dB in PSNR. Additionally, the proposed domain adaptation method significantly improves the base model, especially in severe aberration cases (at most 2.59dB in PSNR). The code and data will be available at https://github.com/zju-jiangqi/OmniLens.