Pith. sign in

REVIEW 3 cited by

Weakly Supervised Lesion Localization With Probabilistic-CAM Pooling

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2005.14480 v1 pith:KGFXHDX3 submitted 2020-05-29 cs.CV

Weakly Supervised Lesion Localization With Probabilistic-CAM Pooling

classification cs.CV
keywords poolinglocalizationpcamlesiongeneratedprobabilisticsupervisiontask
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
Share X Bluesky LinkedIn Reddit HN
read the original abstract

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 leverages the excellent localization ability of CAM during training in a probabilistic fashion. Experiments on the ChestX-ray14 dataset show a ResNet-34 model trained with PCAM pooling outperforms state-of-the-art baselines on both the classification task and the localization task. Visual examination on the probability maps generated by PCAM pooling shows clear and sharp boundaries around lesion regions compared to the localization heatmaps generated by CAM. PCAM pooling is open sourced at https://github.com/jfhealthcare/Chexpert.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 3 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Rethinking Clinical Relevance in Chest X-ray Machine Learning: How Evaluation References Define Performance

    eess.IV 2026-07 conditional novelty 6.0

    Chest X-ray AI model rankings and image-quality metric rankings change substantially with the choice of evaluation reference, so benchmark scores are not neutral.

  2. From Classification to Localization and Clinical Validation: Large-Scale Development of a Deep Learning System for Thoracic Disease Detection on Chest Radiographs in Thailand

    cs.CV 2026-07 conditional novelty 5.0

    A DenseNet-121 + ACM + PCAM system trained on 874,858 Thai chest X-rays reports near-ceiling classification (AUROC 0.994 in-domain, 0.970 cross-site) and 77.9% lesion-localization fraction, with no released code or data.

  3. From Classification to Localization and Clinical Validation: Large-Scale Development of a Deep Learning System for Thoracic Disease Detection on Chest Radiographs in Thailand

    cs.CV 2026-07 conditional novelty 4.0

    Inspectra CXR v5, trained on 874,858 Thai frontal radiographs, reaches mean AUROC 0.994 in-domain and 0.970 across 13 hospitals, with weakly supervised localization LLF 77.9% and radiologist concordance ~94%.