A5005853880- Explainable Artificial Intelligence (XAI)
- AI in cancer detection
- Radiomics and Machine Learning in Medical Imaging
- Cell Image Analysis Techniques
- Image Retrieval and Classification Techniques
- Social Media and Politics
- Advanced Neural Network Applications
- Opinion Dynamics and Social Influence
- Glaucoma and retinal disorders
- Topic Modeling
- Medical Imaging Techniques and Applications
- Complex Network Analysis Techniques
- Maternal and fetal healthcare
- COVID-19 diagnosis using AI
- Retinal Imaging and Analysis
- Retinal Diseases and Treatments
- Ultrasound in Clinical Applications
- Machine Learning and Data Classification
- Misinformation and Its Impacts
- Pregnancy and preeclampsia studies
Technical University of Denmark
2022-2024
Roskilde University
2022
University of Copenhagen
2022
Abstract The placenta is crucial to fetal well-being and it plays a significant role in the pathogenesis of hypertensive pregnancy disorders. Moreover, timely diagnosis previa may save lives. Ultrasound primary imaging modality pregnancy, but high-quality depends on access equipment staff, which not possible all settings. Convolutional neural networks help standardize acquisition images for diagnostics. Our aim was develop deep learning based model classification segmentation ultrasound...
Segmentation uncertainty models predict a distribution over plausible segmentations for given input, which they learn from the annotator variation in training set. However, practice these annotations can differ systematically way are generated, example through use of different labeling tools. This results datasets that contain both data variability and differing label styles. In this paper, we demonstrate applying state-of-the-art segmentation on such lead to model bias caused by We present...
Motivated by the challenging segmentation task of pancreatic tubular networks, this paper tackles two commonly encountered problems in biomedical imaging: Topological consistency segmentation, and expensive or difficult annotation. Our contributions are following: a) We propose a topological score which measures both geometric between predicted ground truth segmentations, applied to model selection validation. b) provide full deep-learning methodology for noisy on time-series image data. In...
Out of distribution (OOD) medical images are frequently encountered, e.g. because site- or scanner differences, image corruption. OOD come with a risk incorrect segmentation, potentially negatively affecting downstream diagnoses treatment. To ensure robustness to such segmentations, we propose Laplacian Segmentation Networks (LSN) that jointly model epistemic (model) and aleatoric (data) uncertainty in segmentation. We capture data spatially correlated logit distribution. For uncertainty,...
Motivated by the challenging segmentation task of pancreatic tubular networks, this paper tackles two commonly encountered problems in biomedical imaging: Topological consistency segmentation, and expensive or difficult annotation. Our contributions are following: a) We propose a topological score which measures both geometric between predicted ground truth segmentations, applied to model selection validation. b) provide full deep-learning methodology for noisy on time-series image data. In...