A Comprehensive Framework for Universal Computational Aberration Correction via Automatic Lens Library Generation and Domain Adaptation
The proposed OmniLens framework provides a robust and flexible solution to universal computational aberration correction, addressing the limitations of lens-specific methods through automatic lens library generation, high-quality codebook priors, and domain adaptation.
Automated Design of Compact Computational Imaging Systems through Global Search Optics
Global Search Optics (GSO) is a comprehensive end-to-end lens design framework that automatically explores the solution space to design compact computational imaging systems with superior imaging quality compared to traditional methods.
Adaptive Cascade Calibrated Multi-Plane Phase Retrieval for Alignment-Free Computational Imaging
A novel Adaptive Cascade Calibrated (ACC) strategy for multi-plane phase retrieval that overcomes misalignment issues through computational self-calibration, enabling alignment-free phase imaging.
Learning Wavefront Modulations to Enhance Imaging Through Scattering Media
This paper introduces a novel end-to-end learning framework to optimize acquisition-time wavefront modulations, which significantly enhance the ability to recover scenes obscured by scattering media.
Phase-Guided Light Field Imaging for High-Resolution 3D Reconstruction
A phase-guided light field (PGLF) algorithm is proposed to significantly improve both the spatial and depth resolutions of 3D imaging using off-the-shelf light field cameras, overcoming the limitations of existing active light field techniques.
Efficient Fourier-Domain Inversion for High Dynamic Range Tomography via the Modulo Radon Transform
A novel Fourier domain algorithm is presented for efficient and stable reconstruction of high dynamic range tomographic images from modulo Radon transform measurements, providing mathematical guarantees and advantages over previous spatial domain approaches.
Efficient Atmospheric Turbulence Mitigation using Deep Learning with Physics-Grounded Modeling
The core message of this work is to develop an efficient deep learning-based method for mitigating atmospheric turbulence in images and videos by carefully integrating insights from classical turbulence mitigation algorithms and leveraging a physics-grounded data synthesis approach.
Efficient 3D Scene Reconstruction through Camera-Sonar Fusion using Gaussian Splatting
Combining camera and sonar data through Gaussian splatting enables significantly better 3D geometry reconstruction and novel view synthesis compared to using camera data alone, especially in small baseline imaging scenarios.
Deep Phase Coded Imaging: Joint Reconstruction of All-in-Focus Images and Depth Maps from a Single Captured Image
A self-supervised method for jointly recovering an all-in-focus image and a pixel-level depth map from a single phase-coded captured image, without requiring any training dataset.
Unsupervised Implicit Neural Representations for General Turbulence Mitigation
NeRT, an unsupervised and physically grounded deep learning method, can effectively mitigate various types of turbulence, including atmospheric and water turbulence, without relying on domain-specific priors or large training datasets.