EUVIMEDEUVIMED entdecken
Crossref

Crucial-Diff: A Unified Diffusion Model for Crucial Image and Annotation Synthesis in Data-Scarce Scenarios

Siyue Yao, Mingjie Sun, Eng Gee Lim, Ran Yi, Baojiang Zhong, Moncef Gabbouj

IEEE Transactions on Image Processing · 2025 · 34 · 7024-7036

DOI: 10.1109/tip.2025.3625380 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Transactions on Image Processing, 2025. Verfasst von Siyue Yao und 5 weiteren Autor:innen. Die Arbeit wurde bisher 4-mal zitiert (Crossref) und nennt 76 Literaturangaben. Gefördert von: Young Scientists Fund of the National Natural Science Foundation of China, Jiangsu Province Foundation for Young Scientists, Open Research Fund of Suzhou Key Lab of Multi-modal Data Fusion and Intelligent Healthcare.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The scarcity of data in various scenarios, such as medical, industry and autonomous driving, leads to model overfitting and dataset imbalance, thus hindering effective detection and segmentation performance. Existing studies employ the generative models to synthesize more training samples to mitigate data scarcity. However, these synthetic samples are repetitive or simplistic and fail to provide "crucial information" that targets the downstream model's weaknesses. Additionally, these methods typically require separate training for different objects, leading to computational inefficiencies. To address these issues, we propose Crucial-Diff, a domain-agnostic framework designed to synthesize crucial samples. Our method integrates two key modules. The Scene Agnostic Feature Extractor (SAFE) utilizes a unified feature extractor to capture target information. The Weakness Aware Sample Miner (WASM) generates hard-to-detect samples using feedback from the detection results of downstream model, which is then fused with the output of SAFE module. Together, our Crucial-Diff framework generates diverse, high-quality training data, achieving a pixel-level AP of 83.63% and an F1-MAX of 78.12% on MVTec. On polyp dataset, Crucial-Diff reaches an mIoU of 81.64% and an mDice of 87.69%. Code is publicly available at https://github.com/JJessicaYao/Crucial-diff.

Förderung

  • Young Scientists Fund of the National Natural Science Foundation of China
  • Jiangsu Province Foundation for Young Scientists
  • Open Research Fund of Suzhou Key Lab of Multi-modal Data Fusion and Intelligent Healthcare
  • Jiangsu Manufacturing Strong Province Construction Special Fund Project Research and Development and Industrialization of Intelligent Service Robots Integratin…

Zitieren

APA

Yao, S., Sun, M., Lim, E. G., Yi, R., Zhong, B., & Gabbouj, M. (2025). Crucial-Diff: A Unified Diffusion Model for Crucial Image and Annotation Synthesis in Data-Scarce Scenarios. IEEE Transactions on Image Processing, 34, 7024-7036. https://doi.org/10.1109/tip.2025.3625380

Vancouver

Yao S, Sun M, Lim EG, Yi R, Zhong B, Gabbouj M. Crucial-Diff: A Unified Diffusion Model for Crucial Image and Annotation Synthesis in Data-Scarce Scenarios. IEEE Transactions on Image Processing. 2025;34:7024-7036. doi:10.1109/tip.2025.3625380

BibTeX
@article{Yao2025,
  author = {Yao, Siyue and Sun, Mingjie and Lim, Eng Gee and Yi, Ran and Zhong, Baojiang and Gabbouj, Moncef},
  title = {Crucial-Diff: A Unified Diffusion Model for Crucial Image and Annotation Synthesis in Data-Scarce Scenarios},
  journal = {IEEE Transactions on Image Processing},
  year = {2025},
  volume = {34},
  pages = {7024--7036},
  doi = {10.1109/tip.2025.3625380}
}
Kostenlos starten →