Two research papers submitted to arXiv detail novel pipelines for interactive lesion segmentation in PET/CT scans, specifically for the AUTOPET V challenge. Both approaches focus on distinguishing between FDG and PSMA tracers and incorporate anatomical context to improve accuracy. One method, TRIAGE, uses a 3D STU-Net with masked autoencoding pre-training and an auxiliary organ segmentation model, while the other builds upon the nnU-Net framework with a two-stage training process including online interactive refinement. AI
IMPACT These advanced segmentation models could improve diagnostic accuracy and efficiency in medical imaging for cancer detection.
RANK_REASON Two research papers submitted to arXiv detail novel pipelines for interactive lesion segmentation in PET/CT scans, specifically for the AUTOPET V challenge.
- arXiv
- AUTOPET V
- fludeoxyglucose (18F)
- glutamate carboxypeptidase II
- Hugging Face
- nnU-Net
- Pablo Lozano Jiménez
- computed tomography
- STU-Net
AI-generated summary · Google Gemini · from 2 sources. How we write summaries →