MirrorWorld: Taming Video Diffusion Models for Mirror Reflection Generation
Summary
MirrorWorld is a reflection-aware video inpainting framework that improves mirror reflection generation in videos by separately modeling semantic content (SRD) and geometric spatial arrangements (GTA), achieving better reflection reconstruction than existing image-based and video inpainting baselines.
View Cached Full Text
Cached at: 08/12/26, 08:22 AM
Paper page - MirrorWorld: Taming Video Diffusion Models for Mirror Reflection Generation
Source: https://huggingface.co/papers/2608.07463
Abstract
MirrorWorld improves video mirror reflection synthesis by separately modeling semantic content associations and geometric spatial arrangements through relation distillation and transformation alignment.
Recent advances invideo diffusion models(VDMs) have enabled high-fidelity video synthesis. However, generating mirror reflections remains challenging because the content within a mirror must remain consistent with the surrounding scene. Existing VDMs are not specifically designed to modelscene-to-mirror relationships, which can lead to reflections with incorrect content or inconsistent spatial arrangements. We observe thatmirror reflection generationinvolves two complementary challenges: determining what scene content should be reflected and how the reflected content should be spatially arranged within the mirror region. Motivated by this observation, we propose MirrorWorld, a reflection-awarevideo inpaintingframework that modelsscene-to-mirror relationshipsduring generation. Specifically, we introduceSemantic Relation Distillation(SRD), which transfers relational information from a frozen visual foundation model to encourage semantic associations between visible scene content and mirror regions. We further proposeGeometric Transformation Alignment(GTA), which learns a transformation that guides the spatial arrangement of reflected content. The two components play complementary roles, with SRD modeling what should be reflected and GTA modeling how it should be arranged. To facilitate research on this problem, we construct a benchmark for videomirror reflection generationby repurposing four existing video mirror datasets into a unified reflection reconstruction task. Experimental results show that MirrorWorld achieves improved reflection reconstruction quality over representative image-based reflection generation methods and strongvideo inpaintingbaselines.
View arXiv pageView PDFProject pageGitHub7Add to collection
Get this paper in your agent:
hf papers read 2608\.07463
Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash
Models citing this paper0
No model linking this paper
Cite arxiv.org/abs/2608.07463 in a model README.md to link it from this page.
Datasets citing this paper0
No dataset linking this paper
Cite arxiv.org/abs/2608.07463 in a dataset README.md to link it from this page.
Spaces citing this paper0
No Space linking this paper
Cite arxiv.org/abs/2608.07463 in a Space README.md to link it from this page.
Collections including this paper0
No Collection including this paper
Add this paper to acollectionto link it from this page.
Similar Articles
Mirror Learning
This paper proposes 'Mirror Learning', a framework for imitation learning from third-person observation that uses a fine-tuned video diffusion model for perspective transformation and an inverse dynamics model to synthesize pseudo first-person expert data, showing that this mirror data alone can train effective policies and improve behavior cloning.
From Synthesis to Removal: Physics-Grounded Reflection Simulation and Diffusion-Based Video Dereflection
This paper presents a closed-loop framework combining physics-grounded reflection simulation, a diffusion-based video dereflection model (S2R-Removal), and a new benchmark (S2R-Bench), achieving state-of-the-art video reflection removal with fast inference.
ReflectWorld-MM: An Entity-Oriented Multimodal Memory System for Open-Ended Video Streams
ReflectWorld-MM is an entity-oriented multimodal memory system for open-ended video streams, using hierarchical long-term memory inspired by human memory theory, achieving state-of-the-art accuracy on six benchmarks.
Latent Spatial Memory for Video World Models
This paper introduces latent spatial memory for video world models, storing 3D scene information directly in diffusion latent space to avoid costly pixel-space reconstruction. The proposed Mirage framework achieves up to 10.57x faster generation and 55x memory reduction while achieving state-of-the-art performance on WorldScore and RealEstate10K.
Wonder: Video World Model Done Better
Wonder is a general-purpose video world model that enables real-time, camera-controllable world exploration from an image or conditional video. It introduces camera conditioning via dense coordinate fields, a sparse attention memory mechanism, and techniques to improve distillation, allowing minute-scale video generation at 16 FPS.