Real2SAM2Real: Generative 3D Caches as Complementary Context for Video Diffusion

• URL: http://arxiv.org/abs/2606.00299v1
• Date: Fri, 29 May 2026 19:28:07 GMT
• Title: Real2SAM2Real: Generative 3D Caches as Complementary Context for Video Diffusion
• Abstract summary: Real2SAM2Real is a framework that leverages 3D lifting models to extract an explicitly editable 3D cache.<n>Our framework overcomes typical breakdowns caused by over-reliance on diffusion priors.<n>Our VDM-tailored 3D cache eradicates perspective ambiguities caused by structural holes and erroneous facades.
• Score: 18.992254109269144
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: While Video Diffusion Models (VDMs) excel at synthesizing high-fidelity videos, enabling precise camera and scene control remains challenging. Existing methods predominantly rely on implicit diffusion priors to generate unobserved regions, inevitably leading to structural collapse during high-dynamic movements or complex occlusions. To address this challenge, we propose Real2SAM2Real, a framework that leverages 3D lifting models (e.g., SAM3D) to extract an explicitly editable 3D cache, serving as a robust geometric scaffold for the VDM. By capturing the entire 3D volume of foreground entities rather than just their visible shells, this cache injects holistic spatial priors into the VDM, providing dependable 3D-aware guidance for complex scene dynamics. To effectively leverage this 3D guidance while preserving pre-trained priors, we design a Soft Spatial-Aligned Injection mechanism alongside a minimally invasive fine-tuning strategy tailored for VDMs. Furthermore, we employ masked normal maps as a cross-modal bridge to construct a 3D-free data curation and perturbation pipeline. Extensive experiments demonstrate that Real2SAM2Real enables precise, decoupled control over both camera trajectories and multi-entity motions. By utilizing the complementary context from generative 3D caches, our framework overcomes typical breakdowns caused by over-reliance on diffusion priors, maintaining exceptional spatiotemporal consistency under large camera shifts and severe occlusions. Crucially, by decoupling geometry from appearance, our VDM-tailored 3D cache eradicates perspective ambiguities caused by structural holes and erroneous facades, as well as misleading cues from reflections and refractions. Project website is available at https://jiayi-wu-leo.github.io/real2sam2real

Related papers

• LooseControlVideo: Directorial Video Control using Spatial Blocking [44.01478894715323]
  We present LooseControlVideo, a framework that enables intuitive and expressive control by using sparse, oriented 3D boxes as a "blocking" proxy.<n>We achieve this by fine-tuning a Wan 2.2 backbone on a video dataset annotated with DNOCS, a novel encoding for 3D size and orientation.<n>Our findings indicate a 1.2x to 3x improvement in Trajectory Error; 2x improvement in Rigid Motion Consistency; and a 1.5x to 2x increase in Occlusion Accuracy.
  arXiv  Detail & Related papers  (2026-06-17T18:32:13Z)
• Think, Act, Build: An Agentic Framework with Vision Language Models for Zero-Shot 3D Visual Grounding [34.1504914582344]
  3D Visual Grounding aims to localize objects in 3D scenes via natural language descriptions.<n>We propose "Think, Act, Build (TAB)", a generative 2D-to-3D reconstruction paradigm operating directly on raw RGB-D streams.<n>To overcome the multi-view coverage deficit caused by strict VLM semantic tracking, we introduce the Semantic-Anchored Geometric Expansion.
  arXiv  Detail & Related papers  (2026-04-01T06:12:16Z)
• I3DM: Implicit 3D-aware Memory Retrieval and Injection for Consistent Video Scene Generation [56.33710337846449]
  I3DM is a novel implicit 3D-aware memory mechanism for consistent video scene generation.<n>At the core of our approach is a 3D-aware memory retrieval strategy.<n>To fully utilize the retrieved historical frames, we introduce a 3D-aligned memory injection module.
  arXiv  Detail & Related papers  (2026-03-24T16:45:40Z)
• Generation Models Know Space: Unleashing Implicit 3D Priors for Scene Understanding [50.098085774845195]
  We propose a paradigm shift by leveraging the implicit spatial prior within large-scale generation models.<n>We introduce GeneEGA-3D (Video Extracted Generative Awareness), a plug-and-play framework that repurposes a pre-trained video diffusion model as a Latent World Simulator.
  arXiv  Detail & Related papers  (2026-03-19T17:59:58Z)
• OneWorld: Taming Scene Generation with 3D Unified Representation Autoencoder [90.8453349494245]
  We present OneWorld, a framework that performs diffusion directly within a coherent 3D representation space.<n>OneWorld generates high-quality 3D scenes with superior cross-view consistency compared to state-of-the-art 2D-based methods.
  arXiv  Detail & Related papers  (2026-03-17T03:43:37Z)
• Controllable Egocentric Video Generation via Occlusion-Aware Sparse 3D Hand Joints [87.13154261503168]
  Motion-controllable video generation is crucial for egocentric applications in virtual reality and embodied AI.<n>Existing methods often struggle to achieve 3D-consistent fine-grained hand articulation.<n>We propose a novel framework that generates egocentric videos from a single reference frame.
  arXiv  Detail & Related papers  (2026-03-12T10:02:23Z)
• Mirage: One-Step Video Diffusion for Photorealistic and Coherent Asset Editing in Driving Scenes [11.72942978943414]
  Vision-centric autonomous driving systems rely on diverse and scalable training data to achieve robust performance.<n>Video object editing offers a promising path for data augmentation.<n>Existing methods often struggle to maintain both high visual fidelity and temporal coherence.<n>Mirage builds upon a text-to-video diffusion prior to ensure temporal consistency across frames.
  arXiv  Detail & Related papers  (2025-12-30T13:40:23Z)
• LiftImage3D: Lifting Any Single Image to 3D Gaussians with Video Generation Priors [107.83398512719981]
  Single-image 3D reconstruction remains a fundamental challenge in computer vision.<n>Recent advances in Latent Video Diffusion Models offer promising 3D priors learned from large-scale video data.<n>We propose LiftImage3D, a framework that effectively releases LVDMs' generative priors while ensuring 3D consistency.
  arXiv  Detail & Related papers  (2024-12-12T18:58:42Z)
• 3DTrajMaster: Mastering 3D Trajectory for Multi-Entity Motion in Video Generation [83.98251722144195]
  Previous methods on controllable video generation primarily leverage 2D control signals to manipulate object motions.<n>We introduce 3DTrajMaster, a robust controller that regulates multi-entity dynamics in 3D space.<n>We show that 3DTrajMaster sets a new state-of-the-art in both accuracy and generalization for controlling multi-entity 3D motions.
  arXiv  Detail & Related papers  (2024-12-10T18:55:13Z)

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.