Loic Le Folgoc

Unsupervised Adaptation of 3D CT Foundation Models for 3D CBCT Segmentation

This work proposes a novel unsupervised domain adaptation (UDA) framework based on redundancy-reducing feature alignment, enabling 3D CBCT segmentation with no target-domain annotations or inference-time adaptation, and demonstrates that this approach consistently outperforms existing pretrained foundation model and UD...