IEEE Trans Med Imaging. 2018 Aug;37(8):1943-1954. doi: 10.1109/TMI.2018.2805821. Epub 2018 Feb 13.
Segmentation of brain tumors from magnetic resonance imaging (MRI) data sets is of great importance for improved diagnosis, growth rate prediction, and treatment planning. However, automating this process is challenging due to the presence of severe partial volume effect and considerable variability in tumor structures, as well as imaging conditions, especially for the gliomas. In this paper, we introduce a new methodology that combines random forests and active contour model for the automated segmentation of the gliomas from multimodal volumetric MR images. Specifically, we employ a feature representations learning strategy to effectively explore both local and contextual information from multimodal images for tissue segmentation by using modality specific random forests as the feature learning kernels. Different levels of the structural information is subsequently integrated into concatenated and connected random forests for gliomas structure inferring. Finally, a novel multiscale patch driven active contour model is exploited to refine the inferred structure by taking advantage of sparse representation techniques. Results reported on public benchmarks reveal that our architecture achieves competitive accuracy compared to the state-of-the-art brain tumor segmentation methods while being computationally efficient.
脑肿瘤的磁共振成像（MRI）数据集分割对于提高诊断、生长速度预测和治疗计划非常重要。然而，由于存在严重的部分容积效应以及肿瘤结构和成像条件的巨大可变性，特别是对于神经胶质瘤，使这个过程自动化具有挑战性。在本文中，我们提出了一种新的方法，将随机森林和主动轮廓模型相结合，用于从多模态容积 MR 图像中自动分割神经胶质瘤。具体来说，我们采用特征表示学习策略，通过使用特定于模态的随机森林作为特征学习核，从多模态图像中有效地探索局部和上下文信息，以进行组织分割。随后，将不同层次的结构信息集成到串联和连接的随机森林中，以推断神经胶质瘤的结构。最后，利用稀疏表示技术，利用新的多尺度斑块驱动主动轮廓模型来细化推断的结构。在公共基准上报告的结果表明，与最先进的脑肿瘤分割方法相比，我们的架构在计算效率方面具有竞争力的准确性。
