Cargando…
Advancing diagnostic performance and clinical usability of neural networks via adversarial training and dual batch normalization
Unmasking the decision making process of machine learning models is essential for implementing diagnostic support systems in clinical practice. Here, we demonstrate that adversarially trained models can significantly enhance the usability of pathology detection as compared to their standard counterp...
| Autores principales: | , , , , , , , , , , |
|---|---|
| Formato: | Online Artículo Texto |
| Lenguaje: | English |
| Publicado: |
Nature Publishing Group UK
2021
|
| Materias: | |
| Acceso en línea: | https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8280105/ https://www.ncbi.nlm.nih.gov/pubmed/34262044 http://dx.doi.org/10.1038/s41467-021-24464-3 |