cs.CVMar 5, 2026

Physics-consistent deep learning for blind aberration recovery in mobile optics

Authors: Kartik Jhawar, Tamo Sancho Miguel Tandoc, Khoo Jun Xuan, Wang Lipo

Organizations: Institute for Digital Molecular Analytics and Science (IDMxS), Nanyang Technological University, Singapore 636921 · School of Electrical and Electronic Engineering, Nanyang Technological University, Singapore 639798

Read PDF
Ask AI
Cite

Copy citation

DOINot available for this paper

Abstract

Mobile photography is often limited by complex, lens-specific optical aberrations. While recent deep learning methods approach this as an end-to-end deblurring task, these "black-box" models lack explicit optical modeling and can hallucinate details. Conversely, classical blind deconvolution remains highly unstable. To bridge this gap, we present Lens2Zernike, a deep learning framework that blindly recovers physical optical parameters from a single blurred image. To the best of our knowledge, no prior work has simultaneously integrated supervision across three distinct optical domains. We introduce a novel physics-consistent strategy that explicitly minimizes errors via direct Zernike coefficient regression (z), differentiable physics constraints encompassing both wavefront and point spread function derivations (p), and auxiliary multi-task spatial map predictions (m). Through an ablation study on a ResNet-18 backbone, we demonstrate that our full multi-task framework (z+p+m) yields a 35% improvement over coefficient-only baselines. Crucially, comparative analysis reveals that our approach outperforms two established deep learning methods from previous literature, achieving significantly lower regression errors. Ultimately, we demonstrate that these recovered physical parameters enable stable non-blind deconvolution, providing substantial in-domain improvement on the patented Institute for Digital Molecular Analytics and Science (IDMxS) Mobile Camera Lens Database for restoring diffraction-limited details from severely aberrated mobile captures.

Figures & tables

Model VariantMAE ( λ\lambda )MSE ( λ2\lambda^{2} )
zz (Baseline Coeff-only)0.001970.001975.82×10−55.82\times 10^{-5}
z+Lwavez+\mathcal{L}_{wave} ( pWp_{W} only)0.001560.001565.56×10−55.56\times 10^{-5}
z+Lpsfz+\mathcal{L}_{psf} ( pPp_{P} only)0.001520.001525.19×10−55.19\times 10^{-5}
z+pz+p (Combined Physics)0.001370.001374.73×10−54.73\times 10^{-5}
z+p+mz+p+m (Full Multi-task)0.001284.20 ×10−5\times 10^{-5}
Table 1: Component Ablation Study (ResNet-18)
Figure 1: Prediction accuracy. Scatter plot of predicted vs. true Zernike coefficients for the test set. The tight correlation confirms the model effectively learns diverse aberration modes.
MethodBackboneMAE ( λ\lambda )
DLAO (2025) [ 8 ]LAPANet0.003240.00324
DLWFS (2019) [ 7 ]Xception0.001730.00173
Ours ( z+p+mz+p+m )ResNet-180.00128
Table 2: Comparison with Literature Baselines
Figure 2: Wavefront fidelity. Visual comparison of Oracle (top) vs. Predicted (middle) wavefront maps for three test cases (columns). The difference maps (bottom) visually highlight structural residuals; quantitatively, the absolute magnitude of these residuals is highly constrained, indicating accurate physical reconstruction.
Figure 3: Restoration performance. Qualitative comparison of deconvolution results. The predicted PSFs successfully recover fine structures and cellular boundaries obscured by lens blur (column 2), achieving clarity and structural fidelity comparable to the Oracle deconvolution (column 3). Highlights indicate regions of significant high-frequency recovery.
More in this topic

Physics-Guided Image Restoration

Physics-guided image restoration uses imaging models and physical priors to recover images degraded by noise, blur, haze, scattering, or undersampling.

See all 52 papers

Explore similar work

CardsList
Oct 13, 2026

A Gradient-Regularized Deep Learning Method for Restoring Image in Myopic Vision

Myopic vision degrades the quality of human vision due to spherical eye defocus aberrations that impair the perception of fine spatial details. To address this limitation, this work proposes a deep learning-based approach to computationally design transformed images that allow individuals with myopia in fixed distance to perceive scenes with enhanced clarity. The method trains a neural network to counteract defocus aberrations modeled using the second-order Zernike polynomial Z20Z_2^0. The approach employs a U-Net architecture trained with a loss function combining reconstruction error minimization with spatial difference structural preservation through ℓ1\ell_1 regularization. Training is carried out exclusively on the KITTI dataset, while testing is performed on DIV2K and ImageNet datasets, demonstrating generalization across diverse image distributions. Under myopic conditions modeled by Z20Z_2^0, baseline performance shows visual degradation with PSNR values ranging from 19.47 to 27.42 dB across datasets and severity levels. The proposed gradient-regularized method achieves consistent improvements, delivering PSNR gains ranging from 1.48 to 7.34 dB and significantly lower perceptual compared to classical baselines.

Kebin Contreras, Luis Toscano-Palomino, Rafael Torres +1

Physics School Univ. Industrial of Santander Bucaramanga, Colombia · Dept. of Computer Science Univ. Industrial of Santander Bucaramanga, Colombia

Image RestorationImage Deblurring
AskPDF
Jun 4, 2026

ErA: Error-Aware Deep Unrolling Network for Single Image Defocus Deblurring

We introduce ErA (Error-Aware Deep Unrolling Network), an end-to-end frame work for single-image defocus deblurring. ErA jointly learns a compact kerne basis and per-pixel weights, while an error-aware term in Augmented Lagrangian unrolling corrects kernel estimation errors via alternating updates and ResUNet denoisers. It achieves state-of-the-art PSNR/SSIM on DPDD, RealDOF, and RTF, and shows strong generalization on CUHK without ground truth.

Tu Vo, Chan Y. Park

KC Machine Learning Lab Seoul, South Korea 06181

Image DeblurringComputational ImagingDeep Unfolding
AskPDF
May 23, 2026

Physen-Noise2Noise: Physics-Guided Self-Supervised Defocus Deblurring with Bias Correction under Low-Light Conditions

Low-light, long-exposure defocus deblurring remains a challenging problem due to the simultaneous presence of severe blur and complex biased noise. Existing methods typically rely on simplified noise assumptions, which limits their effectiveness under realistic imaging conditions. In this work, we propose Physen-Noise2Noise, a self-supervised deblurring framework guided by the physical model of defocus imaging, which leverages noisy multi-frame observations without requiring clean reference images. Unlike conventional Noise2Noise-based approaches that assume zero-mean noise, we derive a frequency-domain constraint inherent to the defocus imaging process and incorporate it into the learning framework via a learnable noise bias parameter. In addition, a multi-frame noisy initialization strategy is introduced to suppress complex biased noise prior to deblurring, providing a more stable starting point for reconstruction. This formulation explicitly models biased noise and enables joint bias correction and high-frequency detail recovery during training. Furthermore, we develop a pretrain-finetune variant to enhance robustness and generalization under challenging noise conditions. Extensive experiments on both simulation and real-world datasets demonstrate that the proposed method consistently outperforms state-of-the-art self-supervised approaches for defocus deblurring in the presence of complex biased noise.

Ziyan Huang, Lang Wu, Hongji Wang +3

School of Mathematics and Statistics, Central South University, Changsha, 410083, People’s Republic of China · Key Laboratory for Micro/Nano Optoelectronic Devices of Ministry of Education, Hunan Provincial Key Laboratory of Low-Dimensional Structural Physics and Devices, School of Physics and Electronics, Hunan University, Changsha 410082, People’s Republic of China

Unsupervised Image RestorationImage DeblurringPhysics-Guided Image RestorationImage Restoration
AskPDF