Segment Anything Model for fetal head-pubic symphysis segmentation in intrapartum ultrasound image analysis

Zihao Zhou, Yaosheng Lu, Jieyun Bai, Víctor M. Campello, Fan Feng, Karim Lekadir

Abstract: Highlights•AoP-SAM framework combines the strengths of CNNs and ViT architectures.•A CNN branch containing feature fusion and multi-scale convolution modules.•Adapter modules within the ViT encoder facilitate efficient fine-tuning.•A cross-branch attention integrates features extracted from CNN and ViT branches.•AoP-SAM is the current SOTA method on the PSFHS Challenge of MICCAI 2023.