Joint Deblurring and 3D Reconstruction for Macrophotography

2025-10-01

AI Summary

Severe defocus blur in macro photography degrades image clarity and hinders high-fidelity multi-view 3D reconstruction. To address this, we propose the first joint deblurring and 3D reconstruction framework specifically designed for macro-scale scenes. Given only a sparse set of multi-view blurred images, our method employs differentiable rendering to formulate a self-supervised optimization objective that simultaneously recovers high-fidelity 3D geometry and spatially varying per-pixel defocus blur kernels. Crucially, it requires neither ground-truth blur kernel annotations nor sharp image supervision—thereby eliminating the reliance of conventional deblurring methods on large-scale paired training data or dense viewpoint sampling. Extensive experiments on both synthetic and real macro datasets demonstrate significant improvements in both deblurring quality and 3D reconstruction accuracy. Our approach establishes a novel paradigm for joint geometry-appearance modeling under macro-scale imaging conditions.

Technology Category

• Computer Vision: 3D Computer Vision
• Machine Learning: Multi-instance/Multi-view Learning
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Macro lens has the advantages of high resolution and large magnification, and 3D modeling of small and detailed objects can provide richer information. However, defocus blur in macrophotography is a long-standing problem that heavily hinders the clear imaging of the captured objects and high-quality 3D reconstruction of them. Traditional image deblurring methods require a large number of images and annotations, and there is currently no multi-view 3D reconstruction method for macrophotography. In this work, we propose a joint deblurring and 3D reconstruction method for macrophotography. Starting from multi-view blurry images captured, we jointly optimize the clear 3D model of the object and the defocus blur kernel of each pixel. The entire framework adopts a differentiable rendering method to self-supervise the optimization of the 3D model and the defocus blur kernel. Extensive experiments show that from a small number of multi-view images, our proposed method can not only achieve high-quality image deblurring but also recover high-fidelity 3D appearance.

Problem

• Solving defocus blur in macrophotography imaging
• Enabling 3D reconstruction from multi-view blurry macro images
• Jointly optimizing clear 3D models and defocus blur kernels

Innovation

• Jointly optimizes clear 3D model and defocus blur kernel
• Uses differentiable rendering for self-supervised optimization
• Recovers high-fidelity 3D appearance from few blurry images