mlsanityData to 7 Oct 2026

Defect Transfer GAN: Diverse Defect Synthesis for Data Augmentation

arXiv 2302.08366 ↗First posted 16 Feb 2023cs.CVcited by 9

Not on the map. Why?

Abstract (arXiv)

Data-hunger and data-imbalance are two major pitfalls in many deep learning approaches. For example, on highly optimized production lines, defective samples are hardly acquired while non-defective samples come almost for free. The defects however often seem to resemble each other, e.g., scratches on different products may only differ in a few characteristics. In this work, we introduce a framework, Defect Transfer GAN (DT-GAN), which learns to represent defect types independent of and across various background products and yet can apply defect-specific styles to generate realistic defective images. An empirical study on the MVTec AD and two additional datasets showcase DT-GAN outperforms state-of-the-art image synthesis methods w.r.t. sample fidelity and diversity in defect generation. We further demonstrate benefits for a critical downstream task in manufacturing -- defect classification. Results show that the augmented data from DT-GAN provides consistent gains even in the few samples regime and reduces the error rate up to 51% compared to both traditional and advanced data augmentation methods.

Key facts. Cited by 9 papers in the map (data through 7 Oct 2026). Builds on 37 of its 50 references that are in the map. First posted on arXiv 16 Feb 2023.

Builds on

37 of 50 references.

References · 37

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2018The Unreasonable Effectiveness of Deep Features as a Perceptual Metriccited by 11,029
  3. 2019EfficientNet: Rethinking Model Scaling for Convolutional Neural Networkscited by 6,415
  4. 2018A Style-Based Generator Architecture for Generative Adversarial Networkscited by 5,986
  5. 2022Generative Adversarial Networkscited by 5,603
  6. 2015Rethinking the Inception Architecture for Computer Visioncited by 5,451
  7. 2017mixup: Beyond Empirical Risk Minimizationcited by 5,155
  8. 2016Image-to-Image Translation with Conditional Adversarial Networkscited by 4,510
  9. 2014Deep Learning Face Attributes in the Wildcited by 3,679
  10. 2019Analyzing and Improving the Image Quality of StyleGANcited by 3,580
  11. 2019CutMix: Regularization Strategy to Train Strong Classifiers with Localizable Featurescited by 2,875
  12. 2016Perceptual Losses for Real-Time Style Transfer and Super-Resolutioncited by 2,678
25 more
  1. 2018Large Scale GAN Training for High Fidelity Natural Image Synthesiscited by 2,609
  2. 2017Arbitrary Style Transfer in Real-time with Adaptive Instance Normalizationcited by 1,998
  3. 2016Identity Mappings in Deep Residual Networkscited by 1,960
  4. 2014Unsupervised Domain Adaptation by Backpropagationcited by 1,906
  5. 2017Improved Regularization of Convolutional Neural Networks with Cutoutcited by 1,623
  6. 2016Context Encoders: Feature Learning by Inpaintingcited by 1,424
  7. 2019Semantic Image Synthesis with Spatially-Adaptive Normalizationcited by 1,295
  8. 2020Training Generative Adversarial Networks with Limited Datacited by 1,184
  9. 2019StarGAN v2: Diverse Image Synthesis for Multiple Domainscited by 1,095
  10. 2017StarGAN: Unified Generative Adversarial Networks for Multi-Domain Image-to-Image Translationcited by 1,036
  11. 2016An overview of gradient descent optimization algorithmscited by 1,021
  12. 2018Multimodal Unsupervised Image-to-Image Translationcited by 795
  13. 2018Free-Form Image Inpainting with Gated Convolutioncited by 594
  14. 2018Generative Image Inpainting with Contextual Attentioncited by 578
  15. 2018Domain Adaptive Faster R-CNN for Object Detection in the Wildcited by 442
  16. 2017The Effectiveness of Data Augmentation in Image Classification using Deep Learningcited by 368
  17. 2020Differentiable Augmentation for Data-Efficient GAN Trainingcited by 361
  18. 2017Toward Multimodal Image-to-Image Translationcited by 295
  19. 2019Large Scale Adversarial Representation Learningcited by 238
  20. 2016Semantic Image Inpainting with Deep Generative Modelscited by 156
  21. 2019Mode Seeking Generative Adversarial Networks for Diverse Image Synthesiscited by 138
  22. 2021Defect-GAN: High-Fidelity Defect Synthesis for Automated Defect Inspectioncited by 77
  23. 2019High-Fidelity Image Generation With Fewer Labelscited by 51
  24. 2018Exemplar Guided Unsupervised Image-to-Image Translation with Semantic Consistencycited by 41
  25. 2019Mask Based Unsupervised Content Transfercited by 8

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2023: 3; 2024: 4; 2025: 1; 2026: 1.

9 papers.

  1. 2023Exploring Plain ViT Reconstruction for Multi-class Unsupervised Anomaly Detectioncited by 47
  2. 2024Learning Feature Inversion for Multi-class Anomaly Detection under General-purpose COCO-AD Benchmarkcited by 15
  3. 2023Improving the Effectiveness of Deep Generative Datacited by 4
  4. 2024Synth4Seg -- Learning Defect Data Synthesis for Defect Segmentation using Bi-level Optimizationcited by 3
  5. 2025"Stones from Other Hills can Polish Jade": Zero-shot Anomaly Image Synthesis via Cross-domain Anomaly Injectioncited by 3
  6. 2024Uncertainty-aware Evaluation of Auxiliary Anomalies with the Expected Anomaly Posteriorcited by 2
  7. 2024SYNOSIS: Image synthesis pipeline for machine vision in metal surface inspectioncited by 2
  8. 2023Generative Model-Driven Synthetic Training Image Generation: An Approach to Cognition in Rail Defect Detectioncited by 1
  9. 2026Scaling Quantum Machine Learning without Tricks: Full-Resolution and Diverse Image Generationcited by 1