SyncFix: Fixing 3D Reconstructions via Multi-View Synchronization

2026-04-13

AI Summary

This work addresses the inconsistency between semantics and geometry in multi-view 3D reconstruction by proposing a diffusion-based joint reconstruction framework. The method formulates reconstruction as a multi-view synchronized implicit bridge matching problem, incorporating cross-view consistency constraints during the denoising process to jointly optimize distorted and clean representations. Trained solely on image pairs, the approach generalizes to an arbitrary number of input views and consistently improves reconstruction quality as more views are added. Notably, it outperforms existing methods even without clean reference images, and achieves significantly higher fidelity when sparse clean references are available. Both qualitative and quantitative evaluations demonstrate the superiority of the proposed method.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Multi-instance/Multi-view Learning
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

We present SyncFix, a framework that enforces cross-view consistency during the diffusion-based refinement of reconstructed scenes. SyncFix formulates refinement as a joint latent bridge matching problem, synchronizing distorted and clean representations across multiple views to fix the semantic and geometric inconsistencies. This means SyncFix learns a joint conditional over multiple views to enforce consistency throughout the denoising trajectory. Our training is done only on image pairs, but it generalizes naturally to an arbitrary number of views during inference. Moreover, reconstruction quality improves with additional views, with diminishing returns at higher view counts. Qualitative and quantitative results demonstrate that SyncFix consistently generates high-quality reconstructions and surpasses current state-of-the-art baselines, even in the absence of clean reference images. SyncFix achieves even higher fidelity when sparse references are available.

Problem

• 3D reconstruction
• multi-view consistency
• semantic inconsistency
• geometric inconsistency
• cross-view synchronization

Innovation

• multi-view synchronization
• diffusion-based refinement
• latent bridge matching
• cross-view consistency
• 3D reconstruction