Published September 8, 2024 | Version v1
Model Open

Feature Extraction for Generative Medical Imaging Evaluation: New Evidence Against an Evolving Trend

Authors/Creators

  • 1. ROR icon The University of Texas MD Anderson Cancer Center

Contributors

  • 1. ROR icon The University of Texas MD Anderson Cancer Center
  • 2. ROR icon Rice University
  • 3. ROR icon Baylor College of Medicine

Description

This archive contains the StyleGAN2 model weights that were used to evaluate the efficacy of Fréchet distances calculated with various backbone architectures in the MICCAI 2024 paper entitled "Feature Extraction for Generative Medical Imaging Evaluation: New Evidence Against an Evolving Trend." The archive contains sixteen models split across four datasets: the Segmentation of the Liver Competition 2007 (SLIVER07), ChestX-ray14, the brain tumor dataset from the Medical Segmentation Decathlon (MSD), and the Automated Cardiac Diagnosis Challenge (ACDC). The models are further split across four data augmentation methods: no augmentation, adaptive discriminator augmentation (ADA), adaptive pseudo augmentation (APA), and differentiable augmentation (DiffAugment). The naming convention for the model weights is "stylegan2-dataset-augmentationType-kimg-fid.pkl" where the FID is the Fréchet distance calculated using an ImageNet-trained InceptionV3 architecture.

Files

Files (5.8 GB)

Name Size
md5:572501f5b32cce07721c43a5eab59705
357.3 MB Download
md5:3cb5f491aa28101a84814e347cd4ac54
357.3 MB Download
md5:610c463e8b8ced34d7c9e52c2367ca73
357.3 MB Download
md5:b84f880da807761159dc771205a3030f
357.3 MB Download
md5:08bd38896121b43efc4ec5ff8fc618f0
381.6 MB Download
md5:19354cc692c0b2524505d9130053e4ae
381.6 MB Download
md5:17eedc7b56f35f0f78ea62f4a9b3dd62
381.5 MB Download
md5:45efc933553bde4f72c998fcb7c43cab
381.5 MB Download
md5:5418693563c8e53cdcacee3f848edc01
357.3 MB Download
md5:5418693563c8e53cdcacee3f848edc01
357.3 MB Download
md5:998f4e32618746177b481b937b007313
357.3 MB Download
md5:c66d7cf0790318b9f12213fbe6229136
357.3 MB Download
md5:6d66686595481df2cbf7978dfd4f0c33
363.9 MB Download
md5:a236c51ec3165940c11491a09bac7b1d
363.9 MB Download
md5:d1325bd55ee86de41ac2c5fd84834828
363.9 MB Download
md5:d900c3f10688cf681f9aaa8dae0f3568
363.9 MB Download

Additional details

Related works

Is part of
Conference paper: arXiv:2311.13717 (arXiv)

Software

Repository URL
https://github.com/mckellwoodland/fid-med-eval
Programming language
Python
Development Status
Active

References

  • Karras, T., et al.: Analyzing and improving the image quality of stylegan. In: Zabih, R., et al. (eds.) CVPR. IEEE (2020)
  • Karras, T., et al.: Training generative adversarial networks with limited data. In: Larochelle, H., et al. (eds.) NeurIPS. vol. 33, pp. 12104–12114. Curran Associates, Inc. (2020)
  • Jiang, L., Dai, B., Wu, W., Loy, C.C.: Deceive d: Adaptive pseudo augmentation for gan training with limited data. In: Ranzato, M. (ed.) NeurIPS. vol. 34, pp. 21655–21667. Curran Associates, Inc. (2021)
  • Zhao, S., Liu, Z., Lin, J., Zhu, J.Y., Han, S.: Differentiable augmentation for dataefficient gan training. In: Larochelle, H., et al. (eds.) NeurIPS. vol. 33, pp. 7559– 7570. Curran Associates, Inc. (2020)
  • Heimann, T., et al.: Comparison and evaluation of methods for liver segmentation from ct datasets. IEEE Trans. Med. Imaging 28(8), 1251–1265 (2009). https://doi.org/10.1109/TMI.2009.2013851
  • Wang, X., et al.: Chestx-ray8: Hospital-scale chest x-ray database and benchmarks on weakly-supervised classification and localization of common thorax diseases. In: Chellappa, R., et al. (eds.) CVPR. IEEE (2017)
  • Antonelli, M., et al.: The medical segmentation decathlon. Nat. Commun. 13(1), 4128 (2022). https://doi.org/10.1038/s41467-022-30695-9
  • Simpson, A.L., et al.: A large annotated medical image dataset for the development and evaluation of segmentation algorithms. arXiv:1902.09063 (2019)