AI RESEARCH
Beyond the Embedding Bottleneck: Adaptive Retrieval-Augmented 3D CT Report Generation
arXiv CS.CV
•
ArXi:2603.15822v1 Announce Type: new Automated radiology report generation from 3D CT volumes often suffers from incomplete pathology coverage. We provide empirical evidence that this limitation stems from a representational bottleneck: contrastive 3D CT embeddings encode discriminative pathology signals, yet exhibit severe dimensional concentration, with as few as 2 effective dimensions out of 512. Corroborating this, scaling the language model yields no measurable improvement, suggesting that the bottleneck lies in the visual representation rather than the generator.