[论文解读] Leveraging a Statistical Shape Model for Efficient Generation of Annotated Training Data: A Case Study on Liver Landmarks Segmentation

Denis Krnjaca, Lorena Krames+1 more|arXiv (Cornell University)|Mar 14, 2026

Medical Imaging and Analysis被引用 0

一句话总结

该论文提出一种基于SSM的流程，从单个人工平均形状标注生成大规模带注释的训练集，以训练3D地标分割网络，在合成数据和MedShapeNet肝脏形状上实现高精度。

ABSTRACT

Anatomical landmark segmentation serves as a critical initial step for robust multimodal registration during computer-assisted interventions. Current approaches predominantly rely on deep learning, which often necessitates the extensive manual generation of annotated datasets. In this paper, we present a novel strategy for creating large annotated datasets using a statistical shape model (SSM) based on a mean shape that is manually labeled only once. We demonstrate the method's efficacy through its application to deep-learning-based anatomical landmark segmentation, specifically targeting the detection of the anterior ridge and the falciform ligament in 3D liver shapes. A specialized deep learning network was trained with 8,800 annotated liver shapes generated by the SSM. The network's performance was evaluated on 500 unseen synthetic SSM shapes, yielding a mean Intersection over Union of 91.4% (87.4% for the anterior ridge and 87.6% for the falciform ligament). Subsequently, the network was applied to clinical patient liver shapes, with qualitative evaluation indicating promising results and highlighting the generalizability of the proposed approach. Our findings suggest that the SSM-based data generation approach alleviates the labor-intensive process of manual labeling while enabling the creation of large annotated training datasets for machine learning. Although our study focuses on liver anatomy, the proposed methodology holds potential for a broad range of applications where annotated training datasets play a pivotal role in developing accurate deep-learning models.

研究动机与目标

• 推动计算机辅助干预中多模态配准的3D解剖地标分割改进。
• 提出利用统计形状模型(SSM)从单个手动标注的平均形状生成大规模带标签的数据集。
• 通过用深度网络对两个肝脏地标（前嵴与镰状韧带）进行分割来演示该方法。
• 评估将标签从平均形状转移到SSM生成的形状，并在合成数据和临床数据上的性能。

提出的方法

• 从48个形状构建肝脏统计形状模型以生成11,500个合成肝脏形状。
• 用两个地标手动标注平均形状；通过顶点索引将标签转移给所有生成的形状。
• 在合成带注释数据上训练SPRIN，一种对旋转不变的3D点云分割网络。
• 将点云降采样到4,096个点，并使用Adam(lr=1e-3)进行训练，批量大小为12，在NVIDIA RTX 2080 TI上进行。
• 在500形状的合成测试集以及MedShapeNet肝脏形状上进行评估并进行定性分析。

Figure 1: Overview of the proposed method employing an SSM for data generation. The mean shape derived from the SSM is manually labeled once (in this example: red: anterior ridge, blue: falciform ligament). Subsequently, labels are transferred to generated shapes using vertex indices. The annotated

实验结果

研究问题

• RQ1 地解剖地标是否能从SSM平均形状稳定地转移到生成的形状？
• RQ2 在SSM生成的注释上训练深度网络是否能在未见的临床肝脏形状上实现鲁棒的地标分割？
• RQ3 模型在合成的SSM数据和外部数据集如MedShapeNet上的表现如何？
• RQ4 SSM基于注释转移在3D形状分割中的局限性与潜在改进是什么？

主要发现

• 基于SSM的标签转移在训练数据的聚合人工标签上达到82.45%的准确率（前嵴87.05%，镰状韧带66.06%）。
• 在合成的SSM测试形状上，SPRIN模型的平均IoU为91.4%（前嵴87.4%，镰状韧带87.6%）。
• 对MedShapeNet形状的定性评估显示出有前景的分割，但观察到一些前嵴标注不完整的情况。
• 结果支持通过SSM生成大规模带标签数据集并转移标签来训练鲁棒的3D地标分割模型的可行性。
• 该方法表明对其他解剖地标和带注释训练数据的模态具有泛化潜力。

Figure 2: Two examples illustrating the aggregated labels annotated by the four individuals (first and third shape) and the corresponding results from the label transfer from the mean shape (second and last shape).