Guangdong Provincial Key Laboratory of Traditional Chinese Medicine Information Technology, Jinan University, Guangzhou, 510632, China.
Guangdong Provincial Key Laboratory of Traditional Chinese Medicine Information Technology, Jinan University, Guangzhou, 510632, China; Auckland Bioengineering Institute, University of Auckland, Auckland, 1010, New Zealand.
Comput Biol Med. 2024 Jun;175:108501. doi: 10.1016/j.compbiomed.2024.108501. Epub 2024 Apr 22.
The segmentation of the fetal head (FH) and pubic symphysis (PS) from intrapartum ultrasound images plays a pivotal role in monitoring labor progression and informing crucial clinical decisions. Achieving real-time segmentation with high accuracy on systems with limited hardware capabilities presents significant challenges. To address these challenges, we propose the real-time segmentation network (RTSeg-Net), a groundbreaking lightweight deep learning model that incorporates innovative distribution shifting convolutional blocks, tokenized multilayer perceptron blocks, and efficient feature fusion blocks. Designed for optimal computational efficiency, RTSeg-Net minimizes resource demand while significantly enhancing segmentation performance. Our comprehensive evaluation on two distinct intrapartum ultrasound image datasets reveals that RTSeg-Net achieves segmentation accuracy on par with more complex state-of-the-art networks, utilizing merely 1.86 M parameters-just 6 % of their hyperparameters-and operating seven times faster, achieving a remarkable rate of 31.13 frames per second on a Jetson Nano, a device known for its limited computing capacity. These achievements underscore RTSeg-Net's potential to provide accurate, real-time segmentation on low-power devices, broadening the scope for its application across various stages of labor. By facilitating real-time, accurate ultrasound image analysis on portable, low-cost devices, RTSeg-Net promises to revolutionize intrapartum monitoring, making sophisticated diagnostic tools accessible to a wider range of healthcare settings.
胎儿头部（FH）和耻骨联合（PS）的分段在分娩期间的超声图像中起着至关重要的作用，可用于监测分娩进展并为关键的临床决策提供信息。在硬件资源有限的系统上实现具有高精度的实时分段具有很大的挑战性。为了解决这些挑战，我们提出了实时分割网络（RTSeg-Net），这是一种开创性的轻量级深度学习模型，它结合了创新的分布转移卷积块、标记多层感知机块和高效的特征融合块。RTSeg-Net 旨在实现最佳的计算效率，最小化资源需求的同时显著提高分割性能。我们在两个不同的分娩期间超声图像数据集上进行的全面评估表明，RTSeg-Net 实现了与更复杂的最先进网络相当的分割精度，仅使用 186 万个参数（仅为其超参数的 6%），速度快 7 倍，在 Jetson Nano 上实现了 31.13 帧/秒的惊人速度，Jetson Nano 是一种以计算能力有限而闻名的设备。这些成果突显了 RTSeg-Net 在低功率设备上提供准确、实时分割的潜力，扩大了其在劳动各个阶段的应用范围。通过在便携式、低成本设备上实现实时、准确的超声图像分析，RTSeg-Net 有望彻底改变分娩监测，使复杂的诊断工具更广泛地应用于各种医疗保健环境。
