Search papers, labs, and topics across Lattice.
Department of Medical Physics and Biomedical Engineering, UCL Hawkes Institute, University College London
5
0
5
Explicitly modeling annotator bias and variability can significantly enhance uncertainty calibration in medical image segmentation without sacrificing accuracy.
Realistic distortion simulation enables a novel framework that significantly enhances prostate DWI correction without the need for costly paired data.
Automated image quality assessment can now be achieved with just five samples, transforming how we tackle the dual-scarcity of annotated clinical MRI data.
T2-weighted MRI alone can outperform standard T2w+DWI in prostate cancer localization, thanks to a clever training scheme that leverages DWI as a "privileged" latent modality.
Prostate cancer diagnosis gets a boost: a new vision foundation model, ProFound, beats task-specific models across 11 clinical tasks after pre-training on a massive multi-institutional MRI dataset.
