ORACLE-CT: Anatomy-Aware Support Pooling for CT Classification
Quick Answer
ORACLE-CT introduces an anatomy-aware aggregation framework for abdominal CT classification, enhancing DINOv3 and I3D-ResNet-121 performance on MERLIN with macro-AUROC improvements from 0.838 to 0.858.
Quick Take
ORACLE-CT introduces an anatomy-aware aggregation framework for abdominal CT classification, enhancing DINOv3 and I3D-ResNet-121 performance on MERLIN with macro-AUROC improvements from 0.838 to 0.858. This model also shows significant gains in external evaluations on Duke-Abdomen and AMOS, improving robustness while linking predictions to anatomical evidence.
Key Points
- ORACLE-CT uses multi-organ segmentation for anatomy-aware pooling in CT scans.
- Support-masked pooling improved DINOv3's macro-AUROC from 0.838 to 0.858.
- I3D-ResNet-121 also saw AUROC gains from 0.829 to 0.848 with ORACLE-CT.
- External evaluations showed DINOv3 improved on Duke-Abdomen from 0.802 to 0.835.
- Pillar-0 gains were primarily from learned attention, with minor benefits from anatomical masking.
Article Content
From source RSS / original summaryarXiv:2606. 05460v1 Announce Type: new Abstract: Abdominal CT disease classification is challenging because each scan is a large 3D volume with many possible findings, while diagnostic evidence is often confined to specific organs or anatomical compartments. Most study-level classifiers aggregate encoder features using anatomy-agnostic pooling or attention, creating a mismatch between localized disease evidence and global evidence aggregation.
We propose ORACLE--CT, an encoder-agnostic anatomy-aware aggregation framework that uses multi-organ segmentation to define label-specific anatomical supports and restrict attention pooling to relevant regions. The framework supports single-organ, multi-organ union, comparative, localized, and global support strategies. We evaluate ORACLE--CT with three encoder families: DINOv3, I3D--ResNet-121, and the radiology-native Pillar--0 encoder.
Models are trained end-to-end on MERLIN and evaluated internally and under frozen external transfer to Duke--Abdomen and AMOS. Compared with global average pooling, support-masked pooling improved MERLIN macro-AUROC/AUPRC from 0. 838/0. 638 to 0. 858/0. 676 for DINOv3 and from 0. 829/0. 617 to 0. 848/0. 659 for I3D--ResNet-121. On harmonized 10-label external evaluation, DINOv3 improved on Duke--Abdomen from 0. 802/0. 628 to 0. 835/0. 683 and on AMOS from 0. 742/0. 313 to 0. 762/0.
350, with similar gains for I3D--ResNet-121. For Pillar--0, most gains came from learned attention, with smaller additional benefit from anatomical masking. ORACLE--CT improves discrimination and external robustness while preserving an auditable link between predictions and anatomical evidence.
Reader Mode unavailable (could not extract clean content).
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from arXiv cs.CV
See more →LLM-Guided ANN Index Optimization for Human-Object Interaction Retrieval
A phase-aware LLM agent optimizes human-object interaction retrieval, outperforming Optuna TPE by 33.3% and VDTuner by 34.2% on the HICO-DET benchmark. This method enhances throughput by 15.3x over UniIR and demonstrates strong transferability across vector database management systems.