Skip to content

StreetDiff: Multi-view Street Scenes Generation via Cross-view Consistent Multi-view Stable Diffusion with Structure Prompts

Sep 2026 · 0 citations · 53 references
Computer Science

TL;DR

The proposed StreetDiff introduces a Panorama--Perspective Synergy design to decouple global layout reasoning from local detail synthesis, and incorporates a Panorama Alignment Module (PAM) that establishes spherical-projection-based attention constraints across views to achieve robust cross-view coherence in challenging urban street scene generation tasks.

Abstract

Multi-view diffusion models have shown strong performance in scenes with strong geometric priors and sparse semantics, such as indoor rooms or simple outdoor environments (e.g., fields, courtyards). However, they often fail to maintain cross-view consistency under camera rotation, especially in structurally complex urban environments. Without explicit modeling of spherical correspondence across views, existing approaches tend to produce object duplication, structural distortion, and layout inconsistency. To address this limitation, we propose StreetDiff, a multi-view diffusion framework that explicitly enforces cross-view alignment during denoising. StreetDiff introduces a Panorama--Perspective Synergy design to decouple global layout reasoning from local detail synthesis, and incorporates a Panorama Alignment Module (PAM) that establishes spherical-projection-based attention constraints across views. By injecting structured alignment constraints without modifying the diffusion backbone, our framework achieves robust cross-view coherence in challenging urban street scene generation tasks. In addition, we construct Street360, a large-scale HDR multi-view urban panorama dataset. Extensive experiments demonstrate that StreetDiff significantly improves structural consistency and visual fidelity compared to prior multi-view diffusion generation methods.

View source

Similar papers

VGGT-Diff: Visual Geometry Meets Diffusion for Sparse-View Novel View Synthesis

This work presents VGGT-Diff, a geometry-routed multi-view diffusion model for sparse-view novel view synthesis, and introduces robust geometry conditioning, combining training-time regularization with inference-time guidance for improved robustness.

Kang-Jie Chen, Xiang-Yu Li, Dong-Bin Zhang et al. · 0 citations
Preprint Sep 2026

GeoVerse: World-Consistent Novel View Synthesis in Geometric Latent Space

GeoVerse is a framework that synthesizes world-consistent novel views by performing generation within the geometric latent space of a pretrained 3D foundation model and injecting appearance priors from a video generative model via a ControlNet-style adapter.

Ke-Rui Ren, Tao Lu, Lin-Ning Xu et al. · 0 citations
Open access Aug 2026

Feed-Forward Multi-view Multi-person Reconstruction with Contrastive Human-Aware 3D Representation

A new top-down paradigm is proposed that maintains a unified, instance-centric human-aware 3D space, enabling simultaneous camera calibration, cross-view association, and human reconstruction via cross-modal contrastive learning and allows correspondence reasoning, semantic aggregation, and instance discrimination to b...

Yuan-Wang Yang, Bu-Zhen Huang, Zong-Xuan Ren et al. · 0 citations
Open access 2026

SSDiffNet: Semantic-Guided and Structure-Consistent Diffusion for Aerial Visible-to-Infrared Image Translation

SSiffNet is proposed, a diffusion-based framework that explicitly enforces semantic alignment and structural consistency for aerial visible-to-infrared image translation and consistently outperforms existing state-of-the-art methods in both visual quality and quantitative metrics.

Zhe Guo, Lan Wei, Rui Luo et al. · 0 citations
Conference Open access Sep 2026

Multi-View Alignment and Denoising via Center-Guided Spectral Diffusion

Center-guided spectral diffusion is proposed, which replaces traditional alignment with generative modeling and outperforms state-of-the-art methods on several datasets and alleviates the noise amplification problem commonly found in traditional alignment methods.

Jia-Yuan Wang, Jie Lian, Yong-Quan Shi et al. · 0 citations

Related blog posts

Microsoft Research Blog Nov 16, 2026

TokenEconomizer: Learning from Experience to Make AI Coding Agents More Efficient

AI coding agents are increasingly help developers fix bugs and build features largely on their own, but every interaction with their models consumes tokens. In each session, an agent sends instructions, tool descriptions, and session history to its model. Most efforts to reduce this cost focus on individual tasks. Yet developers return to the same […] The post TokenEconomizer: Learning from Experience to Make AI Coding Agents More Efficient appeared first on Microsoft Research.

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.