SPECTRE presented at CVPR 2026
Cris Claessens and Christiaan Viviers presented SPECTRE at CVPR 2026 in Denver, Colorado.
SPECTRE (Self-Supervised & Cross-Modal Pretraining for CT Representation Extraction) is a transformer-based foundation model for volumetric CT imaging. It combines geometry-aware modelling with large-scale self-supervised and cross-modal pretraining to learn robust representations for downstream medical-imaging tasks.
The presentation was a great opportunity to share the work with the computer-vision community and discuss how foundation models can help advance scalable, reliable analysis of 3D medical images.
Read the LinkedIn updates from the presenters.
