HoloTime: Taming Video Diffusion Models for Panoramic 4D Scene Generation

2025-04-30

AI Summary

Existing diffusion models struggle to model the complete spatiotemporal evolution of dynamic 3D scenes, hindering the generation of immersive 4D experiences. To address this, we propose Panoramic Animator—the first diffusion-based framework for panoramic video generation driven by a single image or text prompt. It integrates spatiotemporally consistent spatial-depth and temporal-depth estimation to enable an end-to-end pipeline from generated panoramic video to 4D Gaussian splatting reconstruction. We further introduce 360World, the first large-scale, high-fidelity panoramic video dataset specifically designed for 4D scene reconstruction. Our method achieves significant improvements over state-of-the-art approaches on both panoramic video generation and 360° 4D scene reconstruction, enabling high-fidelity, spatiotemporally coherent, and interactive 4D asset synthesis. This work establishes a novel content creation paradigm for VR/AR applications.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Large Multimodal Models (LMMs)
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning

Application Category

• Search and Retrieval-Augmented AI: Personalized, context-aware and across-device search
• Web Mining and Content Analysis: Large pretrained models with web data
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

The rapid advancement of diffusion models holds the promise of revolutionizing the application of VR and AR technologies, which typically require scene-level 4D assets for user experience. Nonetheless, existing diffusion models predominantly concentrate on modeling static 3D scenes or object-level dynamics, constraining their capacity to provide truly immersive experiences. To address this issue, we propose HoloTime, a framework that integrates video diffusion models to generate panoramic videos from a single prompt or reference image, along with a 360-degree 4D scene reconstruction method that seamlessly transforms the generated panoramic video into 4D assets, enabling a fully immersive 4D experience for users. Specifically, to tame video diffusion models for generating high-fidelity panoramic videos, we introduce the 360World dataset, the first comprehensive collection of panoramic videos suitable for downstream 4D scene reconstruction tasks. With this curated dataset, we propose Panoramic Animator, a two-stage image-to-video diffusion model that can convert panoramic images into high-quality panoramic videos. Following this, we present Panoramic Space-Time Reconstruction, which leverages a space-time depth estimation method to transform the generated panoramic videos into 4D point clouds, enabling the optimization of a holistic 4D Gaussian Splatting representation to reconstruct spatially and temporally consistent 4D scenes. To validate the efficacy of our method, we conducted a comparative analysis with existing approaches, revealing its superiority in both panoramic video generation and 4D scene reconstruction. This demonstrates our method's capability to create more engaging and realistic immersive environments, thereby enhancing user experiences in VR and AR applications.

Problem

• Generating panoramic 4D scenes for immersive VR/AR experiences
• Overcoming limitations of static 3D or object-level diffusion models
• Transforming single prompts/images into consistent 4D Gaussian Splatting representations

Innovation

• Integrates video diffusion for panoramic video generation
• Uses 360World dataset for 4D scene reconstruction
• Applies space-time depth estimation for 4D Gaussian Splatting