Corpus ID: 215776849

Weakly Supervised Lesion Localization With Probabilistic-CAM Pooling

@article{Ye2020WeaklySL,
  title={Weakly Supervised Lesion Localization With Probabilistic-CAM Pooling},
  author={W. Ye and J. Yao and H. Xue and Y. Li},
  journal={ArXiv},
  year={2020},
  volume={abs/2005.14480}
}
  • W. Ye, J. Yao, +1 author Y. Li
  • Published 2020
  • Computer Science
  • ArXiv
  • Localizing thoracic diseases on chest X-ray plays a critical role in clinical practices such as diagnosis and treatment planning. However, current deep learning based approaches often require strong supervision, e.g. annotated bounding boxes, for training such systems, which is infeasible to harvest in large-scale. We present Probabilistic Class Activation Map (PCAM) pooling, a novel global pooling operation for lesion localization with only image-level supervision. PCAM pooling explicitly… CONTINUE READING
    4 Citations

    Figures, Tables, and Topics from this paper

    References

    SHOWING 1-10 OF 11 REFERENCES
    Attention-Guided Curriculum Learning for Weakly Supervised Classification and Localization of Thoracic Diseases on Chest Radiographs
    • 55
    • Highly Influential
    • PDF
    Deep multiscale convolutional feature learning for weakly supervised localization of chest pathologies in X-ray images
    • 18
    • Highly Influential
    • PDF
    Weakly Supervised Medical Diagnosis and Localization from Multiple Resolutions
    • 58
    • Highly Influential
    • PDF
    ChestNet: A Deep Neural Network for Classification of Thoracic Diseases on Chest Radiography
    • 31
    • Highly Influential
    • PDF
    From image-level to pixel-level labeling with Convolutional Networks
    • 432
    • Highly Influential
    • PDF
    ChestX-Ray8: Hospital-Scale Chest X-Ray Database and Benchmarks on Weakly-Supervised Classification and Localization of Common Thorax Diseases
    • 1,044
    • Highly Influential
    • PDF
    Learning Deep Features for Discriminative Localization
    • 3,044
    • Highly Influential
    • PDF
    Attention-based Deep Multiple Instance Learning
    • 277
    • Highly Influential
    • PDF
    Deep Residual Learning for Image Recognition
    • 59,713
    • Highly Influential
    • PDF
    A Comparison of Five Multiple Instance Learning Pooling Functions for Sound Event Detection with Weak Labeling
    • 57
    • PDF