Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper introduces SARFA, a novel framework designed to enhance medical image segmentation by addressing the ambiguity often encountered in delineating targets like tumors. By employing probabilistic prompting and optimizing segmentation outputs through a radiomics-driven training objective based on Fréchet Radiomic Distance (FRD) and Direct Preference Optimization (DPO), SARFA generates diverse and clinically relevant masks. Evaluations on CT and MRI benchmarks reveal that SARFA significantly outperforms existing methods, highlighting the importance of radiomic feature alignment in achieving accurate segmentations.
SARFA achieves superior medical image segmentation by aligning radiomic features, outperforming traditional methods in ambiguous cases where expert consensus is often lacking.
The Segment Anything Model (SAM) has demonstrated strong generalizability across a variety of segmentation tasks. However, SAM often struggles in situations where the target to be segmented is ambiguous. This poses a problem in medical imaging, where accurate delineation of targets such as tumors is vital, but even expert radiologists can disagree on the appropriate boundary for a target. Addressing this, we propose SARFA (Segment Anything with Radiomic Feature Alignment), a novel framework for improved medical image segmentation. Via probabilistic prompting, SARFA generates a diverse set of plausible masks for each input image and optimizes them with a radiomics-driven training objective based on Fr\'echet Radiomic Distance (FRD) and Direct Preference Optimization (DPO). By minimizing the FRD between masked predicted and ground truth regions within each image, SARFA encourages segmentation outputs whose anatomical and textural characteristics align with clinically meaningful ground truth representations, without relying solely on pixel-level overlap. Evaluated on computed tomography (CT) and magnetic resonance imaging (MRI) benchmarks, SARFA outperforms existing ambiguous segmentation methods, demonstrating the effectiveness of radiomic feature alignment and DPO-style candidate mask ranking as a training objective. Our code is available at https://github.com/tbwa233/SARFA.