TY - GEN

T1 - A Comprehensive Study of Modern Architectures and Regularization Approaches on CheXpert5000

AU - Ihler, Sontje

AU - Kuhnke, Felix

AU - Spindeldreier, Svenja

PY - 2022/9/15

Y1 - 2022/9/15

N2 - Computer aided diagnosis (CAD) has gained an increased amount of attention in the general research community over the last years as an example of a typical limited data application - with experiments on labeled 100k–200k datasets. Although these datasets are still small compared to natural image datasets like ImageNet1k, ImageNet21k and JFT, they are large for annotated medical datasets, where 1k–10k labeled samples are much more common. There is no baseline on which methods to build on in the low data regime. In this work we bridge this gap by providing an extensive study on medical image classification with limited annotations (5k). We present a study of modern architectures applied to a fixed low data regime of 5000 images on the CheXpert dataset. Conclusively we find that models pretrained on ImageNet21k achieve a higher AUC and larger models require less training steps. All models are quite well calibrated even though we only fine-tuned on 5000 training samples. All ‘modern’ architectures have higher AUC than ResNet50. Regularization of Big Transfer Models with MixUp or Mean Teacher improves calibration, MixUp also improves accuracy. Vision Transformer achieve comparable or on par results to Big Transfer Models.

AB - Computer aided diagnosis (CAD) has gained an increased amount of attention in the general research community over the last years as an example of a typical limited data application - with experiments on labeled 100k–200k datasets. Although these datasets are still small compared to natural image datasets like ImageNet1k, ImageNet21k and JFT, they are large for annotated medical datasets, where 1k–10k labeled samples are much more common. There is no baseline on which methods to build on in the low data regime. In this work we bridge this gap by providing an extensive study on medical image classification with limited annotations (5k). We present a study of modern architectures applied to a fixed low data regime of 5000 images on the CheXpert dataset. Conclusively we find that models pretrained on ImageNet21k achieve a higher AUC and larger models require less training steps. All models are quite well calibrated even though we only fine-tuned on 5000 training samples. All ‘modern’ architectures have higher AUC than ResNet50. Regularization of Big Transfer Models with MixUp or Mean Teacher improves calibration, MixUp also improves accuracy. Vision Transformer achieve comparable or on par results to Big Transfer Models.

KW - Limited data

KW - Medical image classification

KW - Transfer learning

UR - http://www.scopus.com/inward/record.url?scp=85138802602&partnerID=8YFLogxK

U2 - 10.1007/978-3-031-16431-6_62

DO - 10.1007/978-3-031-16431-6_62

M3 - Conference contribution

AN - SCOPUS:85138802602

SN - 9783031164309

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 654

EP - 663

BT - Medical Image Computing and Computer Assisted Intervention – MICCAI 2022 - 25th International Conference, Proceedings

A2 - Wang, Linwei

A2 - Dou, Qi

A2 - Fletcher, P. Thomas

A2 - Speidel, Stefanie

A2 - Li, Shuo

PB - Springer Science and Business Media Deutschland GmbH

T2 - 25th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2022

Y2 - 18 September 2022 through 22 September 2022

ER -