The ability of Segmenting Anything Model (SAM) to segment ultrasound images

• DOI: 10.5582/bst.2023.01128
• Publication Date: 2023-06-21T22:34:55Z
• Authors: Fang Chen; Lingyu Chen; Haojie Han; Sainan Zhang; Daoqiang Zhang; Hongen Liao
• Keywords: Market Segmentation

Accurate ultrasound (US) image segmentation is important for disease screening, diagnosis, and prognosis assessment. However, US images typically have shadow artifacts ambiguous boundaries that affect segmentation. Recently, Segmenting Anything Model (SAM) from Meta AI has demonstrated remarkable potential in a wide range of applications. The purpose this paper was to conduct an initial evaluation the ability SAM segment images, particularly event boundaries. We evaluated SAM's performance on three datasets different tissues, including multi-structure cardiac tissue, thyroid nodules, fetal head. Results indicated generally performs well with clear tissue structures, but it limited Thus, creating improved considers characteristics significant automatic accurate