Dehallu3D: Hallucination-Mitigated 3D Generation from Single Image via Cyclic View Consistency Refinement

2026-03-02

AI Summary

This work addresses the challenge of structural hallucinations—such as spurious holes or protrusions—in single-image 3D reconstruction, which often lead to failures in 3D printing or distortions in virtual environments. To mitigate these geometric artifacts, we propose a cycle-consistency constraint leveraging densely sampled intermediate views, integrated with inter-view consistency and adaptive smoothness control. We introduce the Outlier Risk Measure (ORM), a novel metric to quantitatively assess 3D hallucinations, and design a plug-and-play optimization module that preserves fine details while producing structurally coherent and high-fidelity 3D meshes. Experimental results demonstrate that our approach significantly enhances geometric fidelity, making it well-suited for applications in 3D printing and virtual reality.

Technology Category

• Computer Vision: 3D Computer Vision
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning
• Constraint Satisfaction and Optimization: Other Foundations of Constraint Satisfaction

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web

Abstract

Large 3D reconstruction models have revolutionized the 3D content generation field, enabling broad applications in virtual reality and gaming. Just like other large models, large 3D reconstruction models suffer from hallucinations as well, introducing structural outliers (e.g., odd holes or protrusions) that deviate from the input data. However, unlike other large models, hallucinations in large 3D reconstruction models remain severely underexplored, leading to malformed 3D-printed objects or insufficient immersion in virtual scenes. Such hallucinations majorly originate from that existing methods reconstruct 3D content from sparsely generated multi-view images which suffer from large viewpoint gaps and discontinuities. To mitigate hallucinations by eliminating the outliers, we propose Dehallu3D for 3D mesh generation. Our key idea is to design a balanced multi-view continuity constraint to enforce smooth transitions across dense intermediate viewpoints, while avoiding over-smoothing that could erase sharp geometric features. Therefore, Dehallu3D employs a plug-and-play optimization module with two key constraints: (i) adjacent consistency to ensure geometric continuity across views, and (ii) adaptive smoothness to retain fine details.We further propose the Outlier Risk Measure (ORM) metric to quantify geometric fidelity in 3D generation from the perspective of outliers. Extensive experiments show that Dehallu3D achieves high-fidelity 3D generation by effectively preserving structural details while removing hallucinated outliers.

Problem

• 3D generation
• hallucination
• single image
• structural outliers
• view consistency

Innovation

• hallucination mitigation
• 3D generation
• view consistency
• adaptive smoothness
• outlier risk measure