PostureObjectstitch: Anomaly Image Generation Considering Assembly Relationships in Industrial Scenarios

2026-04-15

AI Summary

Existing methods for industrial anomaly image generation often overlook the assembly posture and orientation relationships among components, resulting in structurally implausible synthetic samples and suboptimal downstream task performance. To address this limitation, this work proposes PostureObjectStitch, a novel approach that explicitly models component assembly relationships within a diffusion framework for the first time. The method disentangles multi-view features into high-frequency, texture, and RGB components, and leverages timestep-aware feature modulation to enable coarse-to-fine consistent generation. Furthermore, it incorporates conditional losses and geometric priors to enforce semantic correctness and spatial plausibility. Experiments on the MureCom benchmark and the newly introduced DreamAssembly dataset demonstrate that the proposed method significantly improves both the visual fidelity of generated images and the performance of downstream anomaly detection tasks.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Machine Learning: Deep Generative Models & Autoencoders

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Image generation technology can synthesize condition-specific images to supplement real-world industrial anomaly data and enhance anomaly detection model performance. Existing generation techniques rarely account for the pose and orientation of industrial components in assembly, making the generated images difficult to utilize for downstream application. To solve this, we propose a novel image synthesis approach, called PostureObjectStitch, that achieves accurate generation to meet the requirement of industrial assembly. A condition decoupling approach is introduced to separate input multi-view images into high-frequency, texture, and RGB features. The feature temporal modulation mechanism adapts these features across diffusion model time-steps, enabling progressive generation from coarse to fine details while maintaining consistency. To ensure semantic accuracy, we introduce a conditional loss that enhances critical industrial elements and a geometric prior that guides component positioning for correct assembly relationships. Comprehensive experimental results on the MureCom dataset, our newly contributed DreamAssembly dataset, and the downstream application validate the outstanding performance of our method.

Problem

• anomaly image generation
• industrial assembly
• component posture
• assembly relationships
• image synthesis

Innovation

• PostureObjectStitch
• condition decoupling
• feature temporal modulation
• geometric prior
• anomaly image generation