From Point to Mask: Interactive Segmentation for 3D Ultrasound

Counts differ because each database indexes a different set of publications. We treat OpenAlex as the canonical count; Google Scholar is not shown (no API, and crawling it violates its ToS).

Segmentation is an important tool in medical imaging but relies on time-consuming manual annotation. Artificial intelligence offers a promising route to support medical professionals, with many models available targeting different modalities and dimensionalities. Most existing work focuses either on 2D images or on three-dimensional CT and MRI, leaving 3D ultrasound (3DUS) largely unexplored. This work evaluates three promptable segmentation models not specifically trained for 3DUS across multiple prompt types and four openly available datasets, to assess their feasibility for this challenging modality.