About this document

Weakly-Supervised Lung Nodule Segmentation Via Training-Free Guidance of 3D Rectified Flow by Anirban Bose is a document available to read on EtoBox.

This paper presents a weakly-supervised segmentation method for lung nodules using a training-free guidance approach that combines pretrained 3D rectified flow models with weakly-supervised predictors. The proposed method enhances segmentation quality by allowing the generative model to operate without retraining, utilizing only image-level labels for fine-tuning. Experiments on the LUNA16 dataset demonstrate significant improvements over traditional methods, indicating the effectiveness of generative found

Author

Anirban Bose

Language

EN