no code implementations • 9 Jun 2023 • Alessandro Wollek, Sardi Hyska, Bastian Sabel, Michael Ingrisch, Tobias Lasser
Deep learning models for image classification are often trained at a resolution of 224 x 224 pixels for historical and efficiency reasons.
no code implementations • 9 Jun 2023 • Alessandro Wollek, Sardi Hyska, Bastian Sabel, Michael Ingrisch, Tobias Lasser
Finally, we propose and evaluate WindowNet, a model that learns optimal window settings, and show that it significantly improves performance compared to the baseline model without windowing.
no code implementations • 9 Jun 2023 • Alessandro Wollek, Philip Haitzer, Thomas Sedlmeyr, Sardi Hyska, Johannes Rueckel, Bastian Sabel, Michael Ingrisch, Tobias Lasser
In this work, we explore the potential of weak supervision of a deep learning-based label prediction model, using a rule-based labeler.
no code implementations • 5 Jun 2023 • Alessandro Wollek, Sardi Hyska, Thomas Sedlmeyr, Philip Haitzer, Johannes Rueckel, Bastian O. Sabel, Michael Ingrisch, Tobias Lasser
This study aimed to develop an algorithm to automatically extract annotations for chest X-ray classification models from German thoracic radiology reports.
1 code implementation • 3 Mar 2023 • Alessandro Wollek, Robert Graf, Saša Čečatka, Nicola Fink, Theresa Willem, Bastian O. Sabel, Tobias Lasser
Conclusion: ViTs performed similarly to CNNs in CXR classification, and their attention-based saliency maps were more useful to radiologists and outperformed GradCAM.
1 code implementation • 1 Aug 2022 • Alessandro Wollek, Theresa Willem, Michael Ingrisch, Bastian Sabel, Tobias Lasser
The proposed IDV approach trained on ID (chest X-ray 14) and OOD data (IRMA and ImageNet) achieved, on average, 0. 999 OOD AUC across the three data sets, surpassing all other OOD detection methods.