#machine learning
Jul 2025
Automatic Fine-grained Segmentation-assisted Report Generation
ASaRG proposes to fuse intermediate features and fine-grained segmentation maps created by specialist radiological models into LLaVA's multi-modal projection layer via simple concatenation, an extension of the popular LLaVA architecture that aims to tackle both of these problems.
F. Jonske, C. Seibold, Osman Alperen Koras et al.
· arXiv.org · 1 citation