Measurement-Oriented Evaluation of Deep Learning Models for Automated Fetal Head Biometry in Prenatal Ultrasound

Short summary

Residual U-Net and TransUNet models achieved high segmentation accuracy (DSC > 0.98, IoU > 0.96) in fetal head ultrasound images, but downstream biometric measurements (HC, BPD, OFD) showed that high segmentation overlap doesn't guarantee superior measurement accuracy.

AI-generated from the title and abstract; the full text is not read.