Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
162 tokens/sec
GPT-4o
7 tokens/sec
Gemini 2.5 Pro Pro
45 tokens/sec
o3 Pro
4 tokens/sec
GPT-4.1 Pro
38 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Dynamic Policy-Driven Adaptive Multi-Instance Learning for Whole Slide Image Classification (2403.07939v1)

Published 9 Mar 2024 in cs.CV

Abstract: Multi-Instance Learning (MIL) has shown impressive performance for histopathology whole slide image (WSI) analysis using bags or pseudo-bags. It involves instance sampling, feature representation, and decision-making. However, existing MIL-based technologies at least suffer from one or more of the following problems: 1) requiring high storage and intensive pre-processing for numerous instances (sampling); 2) potential over-fitting with limited knowledge to predict bag labels (feature representation); 3) pseudo-bag counts and prior biases affect model robustness and generalizability (decision-making). Inspired by clinical diagnostics, using the past sampling instances can facilitate the final WSI analysis, but it is barely explored in prior technologies. To break free these limitations, we integrate the dynamic instance sampling and reinforcement learning into a unified framework to improve the instance selection and feature aggregation, forming a novel Dynamic Policy Instance Selection (DPIS) scheme for better and more credible decision-making. Specifically, the measurement of feature distance and reward function are employed to boost continuous instance sampling. To alleviate the over-fitting, we explore the latent global relations among instances for more robust and discriminative feature representation while establishing reward and punishment mechanisms to correct biases in pseudo-bags using contrastive learning. These strategies form the final Dynamic Policy-Driven Adaptive Multi-Instance Learning (PAMIL) method for WSI tasks. Extensive experiments reveal that our PAMIL method outperforms the state-of-the-art by 3.8\% on CAMELYON16 and 4.4\% on TCGA lung cancer datasets.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (45)
  1. Kingma DP Ba J Adam et al. A method for stochastic optimization. arXiv preprint arXiv:1412.6980, 1412, 2014.
  2. Reinforcement learning based recommender systems: A survey. ACM Computing Surveys, 55(7):1–38, 2022.
  3. Partial policy-based reinforcement learning for anatomical landmark localization in 3d medical images. IEEE Trans. Medical Imaging, 39(4):1245–1255, 2020.
  4. Evaluating reinforcement learning agents for anatomical landmark detection. Medical Image Anal., 53:156–164, 2019.
  5. Diagnostic assessment of deep learning algorithms for detection of lymph node metastases in women with breast cancer. Jama, 318(22):2199–2210, 2017.
  6. Clinical-grade computational pathology using weakly supervised deep learning on whole slide images. Nature medicine, 25(8):1301–1309, 2019.
  7. Histopathology whole slide image analysis with heterogeneous graph representation learning. In CVPR, pages 15661–15670, 2023.
  8. Exploring simple siamese representation learning. In CVPR, pages 15750–15758, 2021.
  9. Rankmix: Data augmentation for weakly supervised learning of classifying whole slide images with diverse sizes and imbalanced categories. In CVPR, pages 23936–23945, 2023.
  10. Imagenet: A large-scale hierarchical image database. In CVPR, pages 248–255. Ieee, 2009.
  11. Multi-scale domain-adversarial multiple-instance cnn for cancer subtype classification with unannotated histopathological images. In CVPR, pages 3852–3861, 2020.
  12. Deep residual learning for image recognition. In CVPR, pages 770–778, 2016.
  13. Reinforcement learning in medical image analysis: Concepts, applications, challenges, and future directions. CoRR, abs/2206.14302, 2022.
  14. Attention-based deep multiple instance learning. In ICML, pages 2127–2136. PMLR, 2018.
  15. Deep multiple instance learning for digital histopathology. In Handbook of Medical Image Computing and Computer Assisted Intervention, pages 521–546. Elsevier, 2020.
  16. Dual-stream multiple instance learning network for whole slide image classification with self-supervised contrastive learning. In CVPR, pages 14318–14328, 2021.
  17. Task-specific fine-tuning via variational information bottleneck for weakly-supervised pathology whole slide image classification. In CVPR, pages 7454–7463, 2023.
  18. Cancer metastasis detection with neural conditional random field. In MIDL, 2018.
  19. Iteratively-refined interactive 3d medical image segmentation with multi-agent reinforcement learning. In CVPR, pages 9394–9402, 2020.
  20. Interventional bag multi-instance learning on whole-slide pathological images. In CVPR. IEEE, 2023.
  21. SGDR: stochastic gradient descent with warm restarts. In ICLR. OpenReview.net, 2017.
  22. Data-efficient and weakly supervised computational pathology on whole-slide images. Nature biomedical engineering, 5(6):555–570, 2021.
  23. Unsupervised reverse domain adaptation for synthetic medical images via adversarial training. IEEE Trans. Medical Imaging, 37(12):2572–2581, 2018.
  24. A framework for multiple-instance learning. Advances in neural information processing systems, 10, 1997.
  25. Learning where to see: A novel attention model for automated immunohistochemical scoring. IEEE Trans. Medical Imaging, 38(11):2620–2631, 2019.
  26. Dgmil: Distribution guided multiple instance learning for whole slide image classification. In MICCAI, pages 24–34. Springer, 2022.
  27. Proximal policy optimization algorithms. arXiv preprint arXiv:1707.06347, 2017.
  28. Bidirectional recurrent neural networks. IEEE transactions on Signal Processing, 45(11):2673–2681, 1997.
  29. Transmil: Transformer based correlated multiple instance learning for whole slide image classification. Advances in Neural Information Processing Systems, 34:2136–2147, 2021.
  30. Lnpl-mil: Learning from noisy pseudo labels for promoting multiple instance learning in whole slide image. In CVPR, pages 21495–21505, 2023.
  31. Reinforcement learning: An introduction. MIT press, 2018.
  32. Multiple instance learning framework with masked hard instance mining for whole slide image classification. In CVPR, pages 4078–4087, 2023.
  33. Review the cancer genome atlas (tcga): an immeasurable source of knowledge. Contemporary Oncology/Współczesna Onkologia, 2015(1):68–77, 2015.
  34. Attention is all you need. Advances in neural information processing systems, 30, 2017.
  35. Diagnose like a pathologist: Transformer-enabled hierarchical attention-guided multiple instance learning for whole slide image classification. arXiv preprint arXiv:2301.08125, 2023.
  36. Searching learning strategy with reinforcement learning for 3d medical image segmentation. In MICCAI, pages 3–11. Springer, 2019.
  37. Remix: A general and efficient framework for multiple instance learning based whole slide image classification. In MICCAI, pages 35–45. Springer, 2022.
  38. Protodiv: Prototype-guided division of consistent pseudo-bags for whole-slide image classification. arXiv preprint arXiv:2304.06652, 2023.
  39. Fusing pre-trained language models with multimodal prompts through reinforcement learning. In CVPR, pages 10845–10856, 2023.
  40. DTFD-MIL: double-tier feature distillation multiple instance learning for histopathology whole slide image classification. In CVPR, pages 18780–18790, 2022.
  41. Multi-level multiple instance learning with transformer for whole slide image classification. arXiv preprint arXiv:2306.05029, 2023.
  42. Continuous-time decision transformer for healthcare applications. In AISTATS, pages 6245–6262. PMLR, 2023.
  43. Rlogist: fast observation strategy on whole-slide images with deep reinforcement learning. In AAAI, volume 37, pages 3570–3578, 2023.
  44. Learning how to detect: A deep reinforcement learning method for whole-slide melanoma histopathology images. Comput. Medical Imaging Graph., 108:102275, 2023.
  45. Murcl: Multi-instance reinforcement contrastive learning for whole slide image classification. IEEE Trans. Medical Imaging, 2022.
Citations (1)
View on Semantic Scholar

Summary

We haven't generated a summary for this paper yet.

Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize Now
X Twitter Logo Streamline Icon: https://streamlinehq.com

Tweets