Abstract
Industrial anomaly inspection faces a major challenge due to the lack of real-world anomaly samples. While generative models are used to create anomaly data, existing methods still struggle when handling small-scale this http URL failure occurs because extreme downsampling in diffusion models causes the information of small anomalies to be lost in the latent this http URL address this, we introduce UniScale, a unified training and inference framework for high-fidelity industrial anomaly generation across arbitrary this http URL training, we introduce an Error-Suppressed Multi-Scale Training (EMT) strategy, which enables the model to learn the rich location-aware textures of anomalies, while suppressing upsampling-induced interpolation errors in texture acquisition, ensuring the model is capable of learning small-scale anomalies, while remaining effective for regular scale this http URL inference, we propose Generation-then-Fusion Denoising. It decouples anomaly generation from background integration, preventing small anomalies from being this http URL experiments demonstrate that our method outperforms state-of-the-art competitors in both anomaly generation quality and downstream detection this http URL achieves a relative IS(a) improvement of 45.86\% (from 1.81 to 2.64) on VisA and 37.70\% (from 1.22 to 1.68) on MVTec AD 2, while also improving the downstream pixel-level IoU by 4.22\% on VisA and AUROC by 6.55\% on MVTec AD 2. Code is available at this https URL.
Abstract (translated)
URL
https://arxiv.org/abs/2608.07864
