de Araújo Adriel Silva, Pinho Márcio Sarroglia, Marques da Silva Ana Maria, Fiorentini Luis Felipe, Becker Jefferson
School of Technology, Pontifícia Universidade Católica do Rio Grande do Sul, Porto Alegre 90619-900, Brazil.
Hospital das Clínicas, Faculdade de Medicina, Universidade de São Paulo, São Paulo 05403-010, Brazil.
J Imaging. 2024 Jul 3;10(7):161. doi: 10.3390/jimaging10070161.
Precise annotations for large medical image datasets can be time-consuming. Additionally, when dealing with volumetric regions of interest, it is typical to apply segmentation techniques on 2D slices, compromising important information for accurately segmenting 3D structures. This study presents a deep learning pipeline that simultaneously tackles both challenges. Firstly, to streamline the annotation process, we employ a semi-automatic segmentation approach using bounding boxes as masks, which is less time-consuming than pixel-level delineation. Subsequently, recursive self-training is utilized to enhance annotation quality. Finally, a 2.5D segmentation technique is adopted, wherein a slice of a volumetric image is segmented using a pseudo-RGB image. The pipeline was applied to segment the carotid artery tree in T1-weighted brain magnetic resonance images. Utilizing 42 volumetric non-contrast T1-weighted brain scans from four datasets, we delineated bounding boxes around the carotid arteries in the axial slices. Pseudo-RGB images were generated from these slices, and recursive segmentation was conducted using a Res-Unet-based neural network architecture. The model's performance was tested on a separate dataset, with ground truth annotations provided by a radiologist. After recursive training, we achieved an Intersection over Union (IoU) score of (0.68 ± 0.08) on the unseen dataset, demonstrating commendable qualitative results.
对大型医学图像数据集进行精确标注可能非常耗时。此外，在处理感兴趣的体积区域时，通常会在二维切片上应用分割技术，这会损失用于准确分割三维结构的重要信息。本研究提出了一种深度学习流程，可同时应对这两个挑战。首先，为了简化标注过程，我们采用了一种半自动分割方法，使用边界框作为掩码，这比像素级描绘耗时更少。随后，利用递归自训练来提高标注质量。最后，采用了一种2.5D分割技术，其中使用伪RGB图像对体积图像的切片进行分割。该流程被应用于在T1加权脑磁共振图像中分割颈动脉树。利用来自四个数据集的42个体积非增强T1加权脑部扫描，我们在轴向切片中勾勒出颈动脉周围的边界框。从这些切片生成伪RGB图像，并使用基于Res-Unet的神经网络架构进行递归分割。该模型的性能在一个单独的数据集上进行了测试，由放射科医生提供了真实标注。经过递归训练后，我们在未见过的数据集上实现了交并比（IoU）分数为（0.68±0.08），展示了值得称赞的定性结果。
