A transfer learning − based DeepLabV3–UNet3+ hybrid model for automated fetal head circumference measurement

Key Points

• Automated measurement of fetal head circumference is significantly enhanced with this hybrid model.
• The model utilizes a transfer learning approach, improving the accuracy of measurements to over 95%.
• Implementation of a deep learning architecture combining DeepLabV3 and UNet3+, optimizing image segmentation performance.
• This method suggests advancements in prenatal imaging but requires external validation for clinical application.