ResNet-34 with Lightweight Decoder for Accurate and Efficient Segmentation of Fetal Brain MRI
Researchers have developed a new deep learning model for segmenting fetal brain MRI scans, aiming to improve prenatal diagnosis. The model combines a ResNet-34 encoder with a lightweight decoder using MLP modules to enhance boundary preservation and reduce errors from motion artifacts. This approach achieves high accuracy, outperforming existing architectures on the FeTA 2021 dataset and demonstrating efficiency suitable for clinical integration. AI
IMPACT Enhances diagnostic capabilities in prenatal care by improving the accuracy and efficiency of fetal brain MRI analysis.