mlsanityData to 7 Oct 2026

CheXtransfer: Performance and Parameter Efficiency of ImageNet Models for Chest X-Ray Interpretation

arXiv 2101.06871 ↗First posted 18 Jan 2021cs.CV cs.AI cs.LGcited by 57

Not on the map. Why?

Abstract (arXiv)

Deep learning methods for chest X-ray interpretation typically rely on pretrained models developed for ImageNet. This paradigm assumes that better ImageNet architectures perform better on chest X-ray tasks and that ImageNet-pretrained weights provide a performance boost over random initialization. In this work, we compare the transfer performance and parameter efficiency of 16 popular convolutional architectures on a large chest X-ray dataset (CheXpert) to investigate these assumptions. First, we find no relationship between ImageNet performance and CheXpert performance for both models without pretraining and models with pretraining. Second, we find that, for models without pretraining, the choice of model family influences performance more than size within a family for medical imaging tasks. Third, we observe that ImageNet pretraining yields a statistically significant boost in performance across architectures, with a higher boost for smaller architectures. Fourth, we examine whether ImageNet architectures are unnecessarily large for CheXpert by truncating final blocks from pretrained models, and find that we can make models 3.25x more parameter-efficient on average without a statistically significant drop in performance. Our work contributes new experimental evidence about the relation of ImageNet to chest x-ray interpretation performance.

Key facts. Cited by 57 papers in the map (data through 7 Oct 2026). Builds on 16 of its 33 references that are in the map. First posted on arXiv 18 Jan 2021.

Builds on

16 of 33 references.

References · 16

  1. 2019PyTorch: An Imperative Style, High-Performance Deep Learning Librarycited by 20,160
  2. 2015Distilling the Knowledge in a Neural Networkcited by 10,076
  3. 2016Xception: Deep Learning with Depthwise Separable Convolutionscited by 2,320
  4. 2019CheXpert: A Large Chest Radiograph Dataset with Uncertainty Labels and Expert Comparisoncited by 1,410
  5. 2017ChestX-ray8: Hospital-scale Chest X-ray Database and Benchmarks on Weakly-Supervised Classification and Localization of Common Thorax Diseasescited by 969
  6. 2018Do Better ImageNet Models Transfer Better?cited by 560
  7. 2017CheXNet: Radiologist-Level Pneumonia Detection on Chest X-Rays with Deep Learningcited by 459
  8. 2018Rethinking ImageNet Pre-trainingcited by 376
  9. 2021Big Self-Supervised Models Advance Medical Image Classificationcited by 320
  10. 2017A Survey of Model Compression and Acceleration for Deep Neural Networkscited by 298
  11. 2014Speeding up Convolutional Neural Networks with Low Rank Expansionscited by 283
  12. 2015Data-free parameter pruning for Deep Neural Networkscited by 112
4 more
  1. 2020Covid-19: Automatic detection from X-Ray images utilizing Transfer Learning with Convolutional Neural Networkscited by 92
  2. 2020Comparing Different Deep Learning Architectures for Classification of Chest Radiographscited by 24
  3. 2020CheXpedition: Investigating Generalization Challenges for Translation of Chest X-Ray Algorithms to the Clinical Settingcited by 18
  4. 2020AutoLR: Layer-wise Pruning and Auto-tuning of Learning Rates in Fine-tuning of Deep Networkscited by 2

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 9; 2022: 15; 2023: 15; 2024: 8; 2025: 7; 2026: 3.

57 papers.

Sort:

By area:

  1. 2021Escaping the Big Data Paradigm with Compact Transformerscited by 261
  2. 2022Benchmarking Self-Supervised Learning on Diverse Pathology Datasetscited by 148
  3. 2022Improving Chest X-Ray Report Generation by Leveraging Warm Startingcited by 93
  4. 2021Causally motivated Shortcut Removal Using Auxiliary Labelscited by 55
  5. 2023Does progress on ImageNet transfer to real-world datasets?cited by 35
  6. 2022Locality Guidance for Improving Vision Transformers on Tiny Datasetscited by 33
  7. 2021How Transferable Are Self-supervised Features in Medical Image Classification Tasks?cited by 32
  8. 2022A Data-Based Perspective on Transfer Learningcited by 32
  9. 2021DeepSMILE: Contrastive self-supervised pre-training benefits MSI and HRD classification directly from H&E whole-slide images in colorectal and breast cancercited by 27
  10. 2022When does Bias Transfer in Transfer Learning?cited by 23
  11. 2022Significantly improving zero-shot X-ray pathology classification via fine-tuning pre-trained image-text encoderscited by 15
  12. 2021Towards Fine-grained Visual Representations by Combining Contrastive Learning with Image Reconstruction and Attention-weighted Poolingcited by 13
38 more
  1. 2021Effect of Pre-Training Scale on Intra- and Inter-Domain Full and Few-Shot Transfer Learning for Natural and Medical X-Ray Chest Imagescited by 11
  2. 2023Enhancing Network Initialization for Medical AI Models Using Large-Scale, Unlabeled Natural Imagescited by 11
  3. 2024Multi-Dataset Multi-Task Learning for COVID-19 Prognosiscited by 11
  4. 2022Fairness and robustness in anti-causal predictioncited by 10
  5. 2022Probabilistic Integration of Object Level Annotations in Chest X-ray Classificationcited by 10
  6. 2022Defending against Reconstruction Attacks through Differentially Private Federated Learning for Classification of Heterogeneous Chest X-Ray Datacited by 8
  7. 2023Chest X-ray Image Classification: A Causal Perspectivecited by 8
  8. 2024Time-to-Event Pretraining for 3D Medical Imagingcited by 7
  9. 2021HistoTransfer: Understanding Transfer Learning for Histopathologycited by 6
  10. 2022Learning to diagnose common thorax diseases on chest radiographs from radiology reports in Vietnamesecited by 5
  11. 2025Chest X-ray Foundation Model with Global and Local Representations Integrationcited by 5
  12. 2021Do we still need ImageNet pre-training in remote sensing scene classification?cited by 4
  13. 2023Video Pretraining Advances 3D Deep Learning on Chest CT Taskscited by 4
  14. 2023Learning with Noisy Labels by Adaptive Gradient-Based Outlier Removalcited by 4
  15. 2024A Bayesian Model Selection Criterion for Selecting Pretraining Checkpointscited by 4
  16. 2022Improving Transfer Learning with a Dual Image and Video Transformer for Multi-label Movie Trailer Genre Classificationcited by 3
  17. 2023Transferring Models Trained on Natural Images to 3D MRI via Position Encoded Slice Modelscited by 3
  18. 2023Deep Reinforcement Learning Framework for Thoracic Diseases Classification via Prior Knowledge Guidancecited by 3
  19. 2023MoCo-Transfer: Investigating out-of-distribution contrastive learning for limited-data domainscited by 3
  20. 2025Benchmarking Foundation Models and Parameter-Efficient Fine-Tuning for Prognosis Prediction in Medical Imagingcited by 3
  21. 2023Revisiting Hidden Representations in Transfer Learning for Medical Imagingcited by 2
  22. 2023BenchMD: A Benchmark for Unified Learning on Medical Images and Sensorscited by 2
  23. 2023Instrumental Variable Learning for Chest X-ray Classificationcited by 2
  24. 2024Improving Factuality of 3D Brain MRI Report Generation with Paired Image-domain Retrieval and Text-domain Augmentationcited by 2
  25. 2025Resolution scaling governs DINOv3 transfer performance in chest radiograph classificationcited by 2
  26. 2026The pretraining domain outweighs the training objective in setting the privacy-utility trade-off of differentially private medical image analysiscited by 2
  27. 2022Interaction of a priori Anatomic Knowledge with Self-Supervised Contrastive Learning in Cardiac Magnetic Resonance Imagingcited by 1
  28. 2022ECG beat classification using machine learning and pre-trained convolutional neural networkscited by 1
  29. 2022Inter-model Interpretability: Self-supervised Models as a Case Studycited by 1
  30. 2023Generalized Cross-domain Multi-label Few-shot Learning for Chest X-rayscited by 1
  31. 2024Primary liver cancer classification from routine tumour biopsy using weakly supervised deep learningcited by 1
  32. 2024Learning Generalized Medical Image Representations through Image-Graph Contrastive Pretrainingcited by 1
  33. 2024Explaining the Model, Protecting Your Data: Revealing and Mitigating the Data Privacy Risks of Post-Hoc Model Explanations via Membership Inferencecited by 1
  34. 2025Item Region-based Style Classification Network (IRSN): A Fashion Style Classifier Based on Domain Knowledge of Fashion Expertscited by 1
  35. 2026Bipartite Mode Matching for Vision Training Set Search from a Hierarchical Data Servercited by 1
  36. 2021Tips and Tricks to Improve CNN-based Chest X-ray Diagnosis: A Surveycited by 0
  37. 2022UnrealNAS: Can We Search Neural Architectures with Unreal Data?cited by 0
  38. 2023GraVIS: Grouping Augmented Views from Independent Sources for Dermatology Analysiscited by 0

50 most cited of 57 All 57 (JSON)