UMI3D: Robust 3D Generation on Unconstrained Multi-Image Inputs via Simultaneous Focus Cross-Attention Routing

• URL: http://arxiv.org/abs/2607.24298v1
• Date: Mon, 27 Jul 2026 11:44:05 GMT
• Title: UMI3D: Robust 3D Generation on Unconstrained Multi-Image Inputs via Simultaneous Focus Cross-Attention Routing
• Abstract summary: Recent 3D foundation models can generate high-quality assets from a single image, but degrade markedly on unconstrained multi-image inputs.<n>We propose UMI3D, a training-free and plug-and-play framework that restructures cross-attention for unconstrained multi-image 3D generation.
• Score: 49.1757838265224
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Recent 3D foundation models can generate high-quality assets from a single image, but degrade markedly on unconstrained multi-image inputs, often producing distorted geometry, over-smoothed textures, and chaotic colors. We argue that this failure stems not from limited model capacity, but from a mismatch between single-image cross-attention and the multi-image setting: existing models lack a principled way to decide which image each 3D voxel should trust at each denoising step. Revisiting recent single-image 3D foundation models, we show that explicitly routing each voxel to its most informative image is sufficient to unlock strong performance on inconsistent multi-image inputs. Based on this observation, we propose UMI3D, a training-free and plug-and-play framework that restructures cross-attention for unconstrained multi-image 3D generation. Its core, Simultaneous Focus Cross-Attention (SFC-Attn), activates all conditioning images at each denoising step while allowing each voxel to focus on the single image that best explains it. To enable this routing, we derive the Voxel Reference Score (VRS), a model-intrinsic metric for voxel--image affinity that requires no external matching, segmentation, or correspondence models. Extensive experiments show that UMI3D unlocks the multi-image potential of single-image 3D generation frameworks across diverse tasks. Project Page: UMI3D-Project.github.io.

Related papers

• Axolotl3D: a Unified Framework for Faithful 3D Shape Completion [1.6934548577260198]
  We present Axolotl3D, a 3D generation model that conditions on images, visibility masks, camera parameters, and a partial point cloud.<n>A unified training strategy synthesizes diverse conditioning regimes from large-scale 3D data, enabling robust cross-modal reasoning.<n>Experiments on Toys4K and OmniObject3D demonstrate state-of-the-art performance under both clean and occluded settings.
  arXiv  Detail & Related papers  (2026-07-22T18:38:20Z)
• DeOcc-1-to-3: 3D De-Occlusion from a Single Image via Self-Supervised Multi-View Diffusion [50.90541069907167]
  We propose DeOcc-1-to-3, an end-to-end framework for occlusion-aware multi-view generation.<n>Our self-supervised training pipeline leverages occluded-unoccluded image pairs and pseudo-ground-truth views to teach the model structure-aware completion and view consistency.
  arXiv  Detail & Related papers  (2025-06-26T17:58:26Z)
• Multi-View Large Reconstruction Model via Geometry-Aware Positional Encoding and Attention [54.66152436050373]
  We propose a Multi-view Large Reconstruction Model (M-LRM) to reconstruct high-quality 3D shapes from multi-views in a 3D-aware manner.<n>Specifically, we introduce a multi-view consistent cross-attention scheme to enable M-LRM to accurately query information from the input images.<n>Compared to previous methods, the proposed M-LRM can generate 3D shapes of high fidelity.
  arXiv  Detail & Related papers  (2024-06-11T18:29:13Z)
• Unique3D: High-Quality and Efficient 3D Mesh Generation from a Single Image [28.759158325097093]
  Unique3D is a novel image-to-3D framework for efficiently generating high-quality 3D meshes from single-view images. Our framework features state-of-the-art generation fidelity and strong generalizability.
  arXiv  Detail & Related papers  (2024-05-30T17:59:54Z)
• MVDiff: Scalable and Flexible Multi-View Diffusion for 3D Object Reconstruction from Single-View [0.0]
  This paper proposes a general framework to generate consistent multi-view images from single image or leveraging scene representation transformer and view-conditioned diffusion model. Our model is able to generate 3D meshes surpassing baselines methods in evaluation metrics, including PSNR, SSIM and LPIPS.
  arXiv  Detail & Related papers  (2024-05-06T22:55:53Z)
• VideoMV: Consistent Multi-View Generation Based on Large Video Generative Model [34.35449902855767]
  Two fundamental questions are what data we use for training and how to ensure multi-view consistency. We propose a dense consistent multi-view generation model that is fine-tuned from off-the-shelf video generative models. Our approach can generate 24 dense views and converges much faster in training than state-of-the-art approaches.
  arXiv  Detail & Related papers  (2024-03-18T17:48:15Z)
• ViewDiff: 3D-Consistent Image Generation with Text-to-Image Models [65.22994156658918]
  We present a method that learns to generate multi-view images in a single denoising process from real-world data. We design an autoregressive generation that renders more 3D-consistent images at any viewpoint.
  arXiv  Detail & Related papers  (2024-03-04T07:57:05Z)
• Beyond First Impressions: Integrating Joint Multi-modal Cues for Comprehensive 3D Representation [72.94143731623117]
  Existing methods simply align 3D representations with single-view 2D images and coarse-grained parent category text. Insufficient synergy neglects the idea that a robust 3D representation should align with the joint vision-language space. We propose a multi-view joint modality modeling approach, termed JM3D, to obtain a unified representation for point cloud, text, and image.
  arXiv  Detail & Related papers  (2023-08-06T01:11:40Z)
• Multi-View Consistent Generative Adversarial Networks for 3D-aware Image Synthesis [48.33860286920389]
  3D-aware image synthesis aims to generate images of objects from multiple views by learning a 3D representation. Existing approaches lack geometry constraints, hence usually fail to generate multi-view consistent images. We propose Multi-View Consistent Generative Adrial Networks (MVCGAN) for high-quality 3D-aware image synthesis with geometry constraints.
  arXiv  Detail & Related papers  (2022-04-13T11:23:09Z)
• Efficient Geometry-aware 3D Generative Adversarial Networks [50.68436093869381]
  Existing 3D GANs are either compute-intensive or make approximations that are not 3D-consistent. In this work, we improve the computational efficiency and image quality of 3D GANs without overly relying on these approximations. We introduce an expressive hybrid explicit-implicit network architecture that synthesizes not only high-resolution multi-view-consistent images in real time but also produces high-quality 3D geometry.
  arXiv  Detail & Related papers  (2021-12-15T08:01:43Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.