6 papers
Look What the Probes Dragged In! Real-World Chest X-ray Shortcuts in MedCLIP
Nikolette Pedersen, Regitze Sydendal, Veronika Cheplygina +1
Vision-language models, such as contrastive language-image pre-training (CLIP)-based approaches, have reached state-of-the-art (SOTA) results in medical artificial intelligence. Ho…
Robustness of transferability estimation metrics for medical imaging
Niclas Claßen, Théo Sourget, Dovile Juodelyte +2
In transfer learning, the choice of source model largely influences the performance on a target dataset. Still, selecting a fitting source remains a challenging task, especially in…
Effect of Demographic Bias on Skin Lesion Classification
Ralf Raumanns, Gerard Schouten, Veronika Cheplygina +1
In this study, we evaluate the performance of skin lesion classification using ResNet-based convolutional models, focusing on the impact of demographic bias in training data, parti…
Dataset Diversity Metrics and Impact on Classification Models
Théo Sourget, Niclas Claßen, Jack Junchi Xu +2
The diversity of training datasets is usually perceived as an important aspect to obtain a robust model. However, the definition of diversity is often not defined or differs across…
Robustness and sex differences in skin cancer detection: logistic regression vs CNNs
Nikolette Pedersen, Regitze Sydendal, Andreas Wulff +3
Deep learning has been reported to achieve high performances in the detection of skin cancer, yet many challenges regarding the reproducibility of results and biases remain. This s…
How I failed machine learning in medical imaging -- shortcomings and recommendations
Gaël Varoquaux, Veronika Cheplygina
Medical imaging is an important research field with many opportunities for improving patients' health. However, there are a number of challenges that are slowing down the progress…