Hoel Kervadec

Prompting with the human-touch: evaluating model-sensitivity of foundation models for musculoskeletal CT segmentation

Caroline Magg, Maaike A. ter Wee, Johannes G. G. Dobbe, Geert J. Streekstra et autres

Promptable Foundation Models (FMs), initially introduced for natural image segmentation, have also revolutionized medical image segmentation. The increasing number of models, along with evaluations varying in datasets, metrics, and compared models, makes direct performance comparison between models difficult and complicates the selection …

Prompting with the human-touch: evaluating model-sensitivity of foundation models for musculoskeletal CT segmentation

Caroline Magg, Maaike A. ter Wee, Johannes G. G. Dobbe, Geert J. Streekstra et autres

Promptable Foundation Models (FMs), initially introduced for natural image segmentation, have also revolutionized medical image segmentation. The increasing number of models, along with evaluations varying in datasets, metrics, and compared models, makes direct performance comparison between models difficult and complicates the selection …

Detecting aberrant p53 immunohistochemical expression patterns in patients with Barrett’s esophagus using artificial intelligence

Michel Botros, Luuk Verheijen, Onno J de Boer, Hans Halfwerk et autres

PurposeImmunohistochemistry (IHC) for the tumor suppressor protein p53 is an adjunct biomarker for Barrett’s esophagus (BE)-related dysplasia classification and risk stratification. Four phenotypic staining patterns are distinguished: wild-type (WT), representing normal staining, and three aberrant patterns: overexpression (OE), null mutation (NM), and …

nl (code pays fourni par la source)

Making sense of TILs : recommendations for morphological assessment of tumour‐infiltrating lymphocytes in gastro‐oesophageal carcinoma

Ylva Weeda, R. Salgado, Filip Van Herpe, Daniel Sur et autres

In the era of immune checkpoint inhibitors for cancers, the need for prognostic biomarkers to identify patients most likely to achieve a durable response has become increasingly more relevant. Tumour-infiltrating lymphocytes (TILs) have gained significant interest, as they can be evaluated using …

nl, au, be, us, ro, de, gb, jp, se, cn, fr, no (code pays fourni par la source)

PaSAL:A Deep Learning Pipeline for Pulmonary Artery-Vein Segmentation and Anatomical Labeling in Thoracic CT

Jasper Eppink, Hoel Kervadec, Julian van Capelleveen, Joost Verhoeff et autres

We present PaSAL, a deep learning pipeline for pulmonary artery-vein segmentation and anatomical labeling in thoracic CT. PaSAL combines an nnU-Net-based binary vessel segmentation model with a graph-based anatomical labeling framework that assigns 19 clinically defined vascular classes. The pipeline integrates vessel …

Spatially continuous dual optimization on compactness function for image segmentation

Lu Lin Tan, Jinhua Xu, Jiangfeng Pan, Jing Yun Yuan et autres

cn, nl, it (code pays fourni par la source)

In-Hoc Concept Representations to Regularise Deep Learning in Medical Imaging

Valentina Corbetta, Floris Six Dijkstra, Hoel Kervadec, Kristoffer Wickstrøm et autres

Deep learning models in medical imaging often achieve strong in-distribution performance but struggle to gener-alise under distribution shifts, frequently relying on spurious correlations instead of clinically meaningful features. We introduce LCRReg, a novel regularisation approach that leverages Latent Concept Representations (LCRs) (e.g., …

nl, no (code pays fourni par la source)

In-hoc Concept Representations to Regularise Deep Learning in Medical Imaging

Valentina Corbetta, Floris Six Dijkstra, Regina G. H. Beets‐Tan, Hoel Kervadec et autres

Deep learning models in medical imaging often achieve strong in-distribution performance but struggle to generalise under distribution shifts, frequently relying on spurious correlations instead of clinically meaningful features. We introduce LCRReg, a novel regularisation approach that leverages Latent Concept Representations (LCRs) (e.g., …

Leveraging Point Annotations in Segmentation Learning with Boundary Loss

Eva Breznik, Hoel Kervadec, Filip Malmberg, Joel Kullberg et autres

se, nl, dk (code pays fourni par la source)

Zero-shot capability of SAM-family models for bone segmentation in CT scans

Caroline Magg, Hoel Kervadec, Clara I. Sánchez

The Segment Anything Model (SAM) and similar models build a family of promptable foundation models (FMs) for image and video segmentation. The object of interest is identified using prompts, such as bounding boxes or points. With these FMs becoming part of medical …

Leveraging point annotations in segmentation learning with boundary loss

Eva Breznik, Hoel Kervadec, Filip Malmberg, Joel Kullberg et autres

This paper investigates the combination of intensity-based distance maps with boundary loss for point-supervised semantic segmentation. By design the boundary loss imposes a stronger penalty on the false positives the farther away from the object they occur. Hence it is intuitively inappropriate …

Nested star-shaped objects segmentation using diameter annotations

Robin Camarasa, Hoel Kervadec, M. Eline Kooi, Jeroen Hendrikse et autres

Most current deep learning based approaches for image segmentation require annotations of large datasets, which limits their application in clinical practice. We observe a mismatch between the voxelwise ground-truth that is required to optimize an objective at a voxel level and the …

nl, dk (code pays fourni par la source)