Computer-Aided Grading of Gliomas Combining Automatic Segmentation and Radiomics
Computer-Aided Grading of Gliomas Combining Automatic Segmentation and Radiomics
复制标题
DOI：
10.1109/tsc.2018.2803826
复制
发表时间：
2018-01-01
影响因子：
7.6
通讯作者：
Qiao, Xu
中科院分区：
文献类型：
--
作者：
Chen, Wei;Liu, Boqiang;Qiao, Xu
云机器人的计算卸载正在学术界和工业界受到广泛关注。然而，当前的解决方案面临挑战：1）传统方法没有考虑网络云机器人（NCR）的特征（例如异构性和机器人协作）； 2）它们无法捕获机器人流工作流程（RSW）中任务的特征（例如，严格的延迟要求和不同的任务语义）； 3）他们没有考虑云机器人的服务质量（QoS）问题。在本文中，我们通过提出一种针对 NCR 的 QoS 感知 RSW 分配算法来解决这些问题，该算法联合优化延迟、能源效率和成本，同时考虑 RSW 和 NCR 的特性。我们首先提出了一种新颖的框架，该框架结合了单个机器人、机器人集群和用于计算卸载的远程云。然后，我们在考虑延迟、能耗和运营成本的同时，制定了 NCR 中 RSW 分配的联合 QoS 优化问题，并表明该问题是 NP 难问题。接下来，我们根据RSW和NCR的特点构建数据流图，并将RSW分配问题转化为混合整数线性规划问题。为了在合理的时间内获得接近最优的解决方案，我们还开发了一种启发式算法。将我们的方法与其他方法进行比较的实验表明，性能显着提高，同时提高了服务质量并减少了执行时间。神经胶质瘤是最常见的原发性脑肿瘤，客观分级对于治疗非常重要。本文提出了一种结合自动分割和放射组学的胶质瘤自动计算机辅助诊断方法，可以提高诊断能力。包含 220 个高级别神经胶质瘤和 54 个低级别神经胶质瘤的 MRI 数据用于评估我们的系统。训练多尺度 3D 卷积神经网络来分割整个肿瘤区域。提取了广泛的放射组学特征，包括一阶特征、形状特征和纹理特征。通过使用具有递归特征消除功能的支持向量机进行特征选择，构建了一个具有 5 倍交叉验证的极端梯度增强分类器的 CAD 系统，用于神经胶质瘤的分级。我们的 CAD 系统对于神经胶质瘤的分级非常有效，准确度为 91.27%，加权宏观精度为 91.27%，加权宏观召回率为 91.27%，加权宏观 F1 分数为 90.64%。这表明所提出的 CAD 系统可以帮助放射科医生对神经胶质瘤进行高精度分级，并具有临床应用的潜力。
Computation offloading for cloud robotics is receiving considerable attention in academic and industrial communities. However, current solutions face challenges: 1) traditional approaches do not consider the characteristics of networked cloud robotics (NCR) (e.g., heterogeneity and robotic cooperation); 2) they fail to capture the characteristics of tasks in a robotic streaming workflow (RSW) (e.g., strict latency requirements and varying task semantics); and 3) they do not consider quality-of-service (QoS) issues for cloud robotics. In this paper, we address these issues by proposing a QoS-aware RSW allocation algorithm for NCR with joint optimization of latency, energy efficiency, and cost, while considering the characteristics of both RSW and NCR. We first propose a novel framework that combines individual robots, robot clusters, and a remote cloud for computation offloading. We then formulate the joint QoS optimization problem for RSW allocation in NCR while considering latency, energy consumption, and operating cost, and show that the problem is NP-hard. Next, we construct a data flow graph based on the characteristics of RSW and NCR, and transform the RSW allocation problem into a mixed-integer linear programming problem. To obtain a near-optimal solution in reasonable time, we also develop a heuristic algorithm. Experiments comparing our approach with others demonstrate significant performance gains, with improved QoS and reduced execution times.Gliomas are the most common primary brain tumors, and the objective grading is of great importance for treatment. This paper presents an automatic computer-aided diagnosis of gliomas that combines automatic segmentation and radiomics, which can improve the diagnostic ability. The MRI data containing 220 high-grade gliomas and 54 low-grade gliomas are used to evaluate our system. A multiscale 3D convolutional neural network is trained to segment whole tumor regions. A wide range of radiomic features including first-order features, shape features, and texture features is extracted. By using support vector machines with recursive feature elimination for feature selection, a CAD system that has an extreme gradient boosting classifier with a 5-fold cross-validation is constructed for the grading of gliomas. Our CAD system is highly effective for the grading of gliomas with an accuracy of 91.27%, a weighted macroprecision of 91.27%, a weighted macrorecall of 91.27%, and a weighted macro-F1 score of 90.64%. This demonstrates that the proposed CAD system can assist radiologists for high accurate grading of gliomas and has the potential for clinical applications.