Genetic approach for Cell-by-Cell dynamic spectrum allocation in the heterogeneous scenario

In this paper,a genetic algorithm (GA) is investigated to deal with cell-by-cell dynamic spectrumallocation (DSA) in the heterogeneous scenario with temporal and spatial traffic demand changes,whichis also known as a difficult combinatorial optimization problem.A new two-dimensional chromosome encodingscheme is defined according to characteristics of the heterogeneous scenario,which prevents forminginvalid solutions during the genetic operation and enables much faster convergence.A novel randomcoloring gene generation function is presented which is the basic operation for initialization and mutationin the genetic algorithm.Simulative comparison demonstrates that the proposed GA-based cell-by-cellDSA outperforms the conventional contiguous DSA scheme both in terms of spectral efficiency gain andquality of service (QoS) satisfaction.

融合多层级特征的脑肿瘤图像分割方法

针对脑肿瘤图像分割中网络模型信息损耗、上下文信息联系不足及网络泛化能力较差导致分割精度较低的问题,提出了一种新型的脑肿瘤图像分割方法,该方法是通过深度门控卷积模块(depth gate convolution,DGC)和特征增强模块(feature enhancement module,FEM)组成的多层级连接(multi-level connection,MC)脑肿瘤分割模型。采用深度卷积模块降低特征信息在逐层传递的信息损耗;使用控制门单元(control gate unit,CGU)实现各个尺度的特征图的MC,其中组合池化来减少下采样过程中的信息丢失;通过FEM增强分割区域的特征权重。实验结果表明,预测分割脑肿瘤的整体肿瘤区(whole tumor,WT)、核心肿瘤区(tumor core,TC)和增强肿瘤区(enhancement tumor,ET)的Dice系数分别达到了0.92、0.84和0.83,Hausdorff距离达到了0.77、1.50和0.92,脑肿瘤分割精度相较于当前较多方法分割精度和计算效率较高,具有良好的分割性能。