CrackForward: Context-Aware Severity Stage Crack Synthesis for Data Augmentation

arXiv cs.CV / 4/23/2026

📰 NewsTools & Practical UsageModels & Research

共有:

Key Points

The paper introduces CrackForward, a context-aware generative framework to synthesize realistic crack growth patterns for augmenting scarce, well-annotated crack detection/segmentation datasets.
Unlike approaches that mainly alter textures or backgrounds, CrackForward explicitly models crack morphology by combining directionally guided elongation with learned thickening and branching.
It contributes two main innovations: a context-guided crack expansion module using local directional cues and adaptive random walks, and a two-stage U-Net-style generator to learn spatially varying crack properties.
Experiments indicate the synthetic cracks maintain key characteristics (e.g., stage saturation and thickness) and improve the performance of multiple crack segmentation architectures, suggesting structure-aware synthetic generation is more informative than conventional augmentation.
CrackForward is presented as a new arXiv submission (v1), providing an actionable research direction for structural health monitoring data augmentation.

Abstract

Reliable crack detection and segmentation are vital for structural health monitoring, yet the scarcity of well-annotated data constitutes a major challenge. To address this limitation, we propose a novel context-aware generative framework designed to synthesize realistic crack growth patterns for data augmentation. Unlike existing methods that primarily manipulate textures or background content, CrackForward explicitly models crack morphology by combining directional crack elongation with learned thickening and branching. Our framework integrates two key innovations: (i) a contextually guided crack expansion module, which uses local directional cues and adaptive random walk to simulate realistic propagation paths; and (ii) a two-stage U-Net-style generator that learns to reproduce spatially varying crack characteristics such as thickness, branching, and growth. Experimental results show that the generated samples preserve target-stage saturation and thickness characteristics and improve the performance of several crack segmentation architectures. These results indicate that structure-aware synthetic crack generation can provide more informative training data than conventional augmentation alone.