arXiv 2601.008392025-12-27cs.CV

针对心脏超声的深度分割架构的统一回顾与基准测试：CAMUS

Unified Review and Benchmark of Deep Segmentation Architectures for Cardiac Ultrasound on CAMUS

Department of Computer Science；Artificial Intelligence, Dongguk University, Seoul 04620, Republic of Korea；Department of Semiconductor System Engineering, Sejong University, Seoul, 05006, Republic of Korea；Department of Operations Research；Statistics, Faculty of Organizational Sciences, University of Belgrade, Belgrade, Serbia；Department of Industrial Engineering \& Management, Yuan Ze University, Taoyuan City 320315, Taiwan；Department of Applied Mathematical Science, College of Science；Technology, Korea University, Sejong 30019, Republic of Korea

Zahid Ullah, Muhammad Hilal, Eunsoo Lee, Dragan Pamucar, Jihie Kim

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

本文通过统一基准测试评估了三种心脏超声分割架构的性能，提出标准化预处理方法，并展望了自监督和多模态标注技术的应用。

摘要：

本文结合对心脏超声分割文献的聚焦回顾和三种有影响力的架构（U-Net、Attention U-Net 和 TransUNet）在 Cardiac Acquisitions for Multi-Structure Ultrasound Segmentation（CAMUS）超声心动图数据集上的受控比较。我们的基准测试涵盖了多种预处理路线，包括原生 NIfTI 体积、16 位 PNG 导出、GPT 协助的基于多边形的伪标签以及在数千个未标记的 cine 帧上进行自监督预训练（SSL）。使用相同的训练分割、损失和评估标准，U-Net 在直接训练 NIfTI 数据（保留原生动态范围）时实现了 94% 的平均 Dice 分数，而 PNG-16 位工作流程在类似条件下达到了 91%。Attention U-Net 在小或低对比度区域提供了适度的改进，减少了边界泄漏，而 TransUNet 由于其对全局空间上下文建模的能力，在挑战性帧上表现出最强的泛化能力，尤其是在使用 SSL 初始化时。伪标签扩展了训练集，并在置信度过滤后提高了鲁棒性。总体而言，我们的贡献有三方面：在标准化的 CAMUS 预处理和评估下的 U-Net、Attention U-Net 和 TransUNet 的统一基准测试；在准备超声数据时保持强度保真度、分辨率一致性及对齐的实用指导；以及对可扩展的自监督和新兴的基于 GPT 的多模态标注流程的展望，用于快速标注、质量保证和目标数据集整理。

英文摘要：

Several review papers summarize cardiac imaging and DL advances, few works connect this overview to a unified and reproducible experimental benchmark. In this study, we combine a focused review of cardiac ultrasound segmentation literature with a controlled comparison of three influential architectures, U-Net, Attention U-Net, and TransUNet, on the Cardiac Acquisitions for Multi-Structure Ultrasound Segmentation (CAMUS) echocardiography dataset. Our benchmark spans multiple preprocessing routes, including native NIfTI volumes, 16-bit PNG exports, GPT-assisted polygon-based pseudo-labels, and self-supervised pretraining (SSL) on thousands of unlabeled cine frames. Using identical training splits, losses, and evaluation criteria, a plain U-Net achieved a 94% mean Dice when trained directly on NIfTI data (preserving native dynamic range), while the PNG-16-bit workflow reached 91% under similar conditions. Attention U-Net provided modest improvements on small or low-contrast regions, reducing boundary leakage, whereas TransUNet demonstrated the strongest generalization on challenging frames due to its ability to model global spatial context, particularly when initialized with SSL. Pseudo-labeling expanded the training set and improved robustness after confidence filtering. Overall, our contributions are threefold: a harmonized, apples-to-apples benchmark of U-Net, Attention U-Net, and TransUNet under standardized CAMUS preprocessing and evaluation; practical guidance on maintaining intensity fidelity, resolution consistency, and alignment when preparing ultrasound data; and an outlook on scalable self-supervision and emerging multimodal GPT-based annotation pipelines for rapid labeling, quality assurance, and targeted dataset curation.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。