FetSAM represents a cutting-edge deep learning model aimed at revolutionizing fetal head ultrasound segmentation, thereby elevating prenatal diagnostic precision. Utilizing a comprehensive dataset-the largest to date for fetal head metrics-FetSAM incorporates prompt-based learning. It distinguishes itself with a dual loss mechanism, combining Weighted DiceLoss and Weighted Lovasz Loss, optimized through AdamW and underscored by class weight adjustments for better segmentation balance. Performance benchmarks against prominent models such as U-Net, DeepLabV3, and Segformer highlight its efficacy. FetSAM delivers unparalleled segmentation accuracy, demonstrated by a DSC of 0.90117, HD of 1.86484, and ASD of 0.46645. FetSAM sets a new benchmark in AI-enhanced prenatal ultrasound analysis, providing a robust, precise tool for clinical applications and pushing the envelope of prenatal care with its groundbreaking dataset and segmentation capabilities.

Download full-text PDF

Source
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC11100952PMC
http://dx.doi.org/10.1109/OJEMB.2024.3382487DOI Listing

Publication Analysis

Top Keywords

fetal head
12
segmentation
5
fetsam
4
fetsam advanced
4
advanced segmentation
4
segmentation techniques
4
techniques fetal
4
head biometrics
4
biometrics ultrasound
4
ultrasound imagery
4

Similar Publications

Background: Diprosopus is one of the rarest anomalies. It typically manifests as bilateral alterations and often involves anomalies within the cranial structures. In this report, we present a case of a fetus with diprosopus diagnosed prenatally.

View Article and Find Full Text PDF

Rationale: Microcephaly, epilepsy, and developmental delay (MCSZ) is a rare neurodevelopmental disorder associated with autosomal recessive inheritance of mutations in the polynucleotide kinase 3'-phosphatase (PNKP) gene. Prompt identification and management are essential, as delayed diagnosis or intervention may result in severe complications or mortality. In this case, prenatal screening in the second trimester detected fetal microcephaly with a gradual decline in head circumference, prompting the decision to terminate the pregnancy.

View Article and Find Full Text PDF

Two minutes is sufficient to characterize the viscoelastic properties of the human lower birth canal during the first stage of labor.

Clin Biomech (Bristol)

January 2025

Department of Mechanical Engineering, University of Michigan, 2350 Hayward St, Ann Arbor, MI 48109, United States; Department of Biomedical Engineering, University of Michigan, 200 Bonisteel Blvd, Ann Arbor, MI 48109, United States. Electronic address:

Background: The lower birth canal is the final constriction through which a fetal head must pass for delivery. Unfortunately, injuries to the lower birth canal tissues occur in up to 19 % of first-time vaginal deliveries due to the 300 % stretch required.

Methods: This is a secondary analysis of data from 56 healthy nullipara recorded by a lower birth canal dilator during the first stage of labor.

View Article and Find Full Text PDF

Determining the differential diagnosis of small scalp cysts identified on a fetus is difficult. In particular, many physicians have difficulty differentiating small meningoceles from small scalp cysts during the prenatal period. Volume contrast imaging increases contrast between tissues, thereby allowing an enhanced view of target structures.

View Article and Find Full Text PDF

Aims: Ultrasound is used in the delivery room to assess fetal head position, engagement during labor, and anal sphincter injuries in the immediate postpartum period. The transperineal approach allows for direct visualization of the structures of interest without altering anatomical landmarks. Various ultrasound measurements during labor have been described in the literature, and their use varies widely across maternity units.

View Article and Find Full Text PDF

Want AI Summaries of new PubMed Abstracts delivered to your In-box?

Enter search terms and have AI summaries delivered each week - change queries or unsubscribe any time!