Knowledge-Embedded and Hypernetwork-Guided Few-Shot Substation Meter Defect Image Generation Method
Jackie Alex, Justin Petter
TL;DR
The paper tackles the challenge of generating realistic substation meter defect images under few-shot conditions by integrating three components: knowledge embedding to bridge the domain gap, geometric crack modeling to impose precise defect constraints, and a hypernetwork-guided diffusion process to balance fidelity and controllability. It leverages DreamBooth-style fine-tuning, edge/SAM-based region constraints, and multi-level control signals within Stable Diffusion to produce diverse, region-accurate defect images. Experimental results on the Substation Meter Dataset demonstrate improved generation quality and meaningful gains in downstream defect detection performance, validating the practical utility of synthetic data for industrial inspection. The work offers a scalable data-synthesis solution for industrial contexts with scarce defect samples, with potential extensions to other defect types and temporal data.
Abstract
Substation meters play a critical role in monitoring and ensuring the stable operation of power grids, yet their detection of cracks and other physical defects is often hampered by a severe scarcity of annotated samples. To address this few-shot generation challenge, we propose a novel framework that integrates Knowledge Embedding and Hypernetwork-Guided Conditional Control into a Stable Diffusion pipeline, enabling realistic and controllable synthesis of defect images from limited data. First, we bridge the substantial domain gap between natural-image pre-trained models and industrial equipment by fine-tuning a Stable Diffusion backbone using DreamBooth-style knowledge embedding. This process encodes the unique structural and textural priors of substation meters, ensuring generated images retain authentic meter characteristics. Second, we introduce a geometric crack modeling module that parameterizes defect attributes--such as location, length, curvature, and branching pattern--to produce spatially constrained control maps. These maps provide precise, pixel-level guidance during generation. Third, we design a lightweight hypernetwork that dynamically modulates the denoising process of the diffusion model in response to the control maps and high-level defect descriptors, achieving a flexible balance between generation fidelity and controllability. Extensive experiments on a real-world substation meter dataset demonstrate that our method substantially outperforms existing augmentation and generation baselines. It reduces Frechet Inception Distance (FID) by 32.7%, increases diversity metrics, and--most importantly--boosts the mAP of a downstream defect detector by 15.3% when trained on augmented data. The framework offers a practical, high-quality data synthesis solution for industrial inspection systems where defect samples are rare.