HSM: Hierarchical Scene Motifs for Multi-Scale Indoor Scene Generation

2025-03-21

AI Summary

Existing indoor 3D scene layout generation methods focus predominantly on large furniture, neglecting small objects—leading to sparse, geometrically distorted scenes that fail to satisfy dense spatial arrangements specified in text descriptions. To address this, we propose the Hierarchical Scene Primitive (HSP) framework, the first to explicitly model cross-scale spatial dependencies between surfaces and objects, as well as recurring layout patterns across scales, enabling coherent multi-scale generation—from floor-level furniture to tabletop small objects. Our approach integrates a hierarchical graph neural network with a conditional diffusion model to jointly encode surface semantics, geometric constraints, and multi-granularity layout priors. Evaluated across diverse room types and layout configurations, our method achieves significant improvements in visual plausibility and text-layout alignment over state-of-the-art approaches.

Technology Category

• Natural Language Processing: Language Grounding & Multi-modal NLP
• Computer Vision: 3D Computer Vision
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Semantics and Knowledge: Scalable techniques for the creation, curation, publication, maintenance, and consumption of large, Web-based, structured, reusable, knowledge graphs and ontologies
• Search and Retrieval-Augmented AI: Efficiency and scalability of Web search engines

Abstract

Despite advances in indoor 3D scene layout generation, synthesizing scenes with dense object arrangements remains challenging. Existing methods primarily focus on large furniture while neglecting smaller objects, resulting in unrealistically empty scenes. Those that place small objects typically do not honor arrangement specifications, resulting in largely random placement not following the text description. We present HSM, a hierarchical framework for indoor scene generation with dense object arrangements across spatial scales. Indoor scenes are inherently hierarchical, with surfaces supporting objects at different scales, from large furniture on floors to smaller objects on tables and shelves. HSM embraces this hierarchy and exploits recurring cross-scale spatial patterns to generate complex and realistic indoor scenes in a unified manner. Our experiments show that HSM outperforms existing methods by generating scenes that are more realistic and better conform to user input across room types and spatial configurations.

Problem

• Generating dense indoor 3D scenes with realistic object arrangements
• Addressing neglect of small objects in current scene generation methods
• Ensuring object placement follows hierarchical and textual specifications

Innovation

• Hierarchical framework for multi-scale scene generation
• Exploits recurring cross-scale spatial patterns
• Unified generation of complex realistic indoor scenes