Implementation and decoding method of OCC system based on MIMO

The application of high-frame-rate cameras as well as the complex image processing techniques will lead to a series of problems, such as high system cost and long transmission delay. In this paper, we introduce narrow-band filtering technology to reduce the impact of optical noise and reduce the complexity of image processing from the physical level. We also introduce multiple-input multiple-output (MIMO) technology into the optical camera communication (OCC) system to increase system transmission rate, and propose a light emitting diode (LED) array decoding algorithm based on the directional projection method to reduce the system delay. By accumulating the target pixel values in each row and column of the image, the proposed method then determines the position and boundary of the detected target to distinguish the target area from the background. Experimental results indicate that the communication distance can reach up to 5.5 m without error bits detected. When the LED array at the transmitter of this system flashes at a frequency of 12 Hz, the transmission rate can reach 126.32 bit/s.     

一种新型单层递归神经网络解决非光滑伪凸优化问题

非光滑伪凸优化问题是一类比较特殊的非凸优化问题,常出现在各类科学与工程应用中,因此具有很大的研究价值.针对现有神经网络模型解决非光滑伪凸优化问题存在的不足,该文基于微分包含理论,提出一种新型单层递归神经网络模型.通过理论分析,证明了神经网络状态解在有限时间内收敛到可行域,且永驻其中,最终神经网络状态解收敛于原优化问题的最优解.最后,通过数值实验,验证了所提理论的有效性.与现有的神经网络相比,该文所提神经网络模型结构简单仅为单层,不需要提前计算罚参数,且对初始点选取没有任何特殊的要求.     

基于多流网络一致性的视频显著性检测

现有的视频显著性检测算法通常采用双流结构提取视频的时空线索,其中运动信息作为双流结构的一个分支,在显著物体发生剧烈或慢速移动时存在运动估计准确率低的问题,并且不合理的训练数据或方案使得权重偏向单个分支结构。提出一种基于多流网络一致性的视频显著性检测算法MSNC。设计并使用一种新的三重网络结构提取预选目标区域的颜色信息、时序信息和先验特征,通过先验特征补偿运动流的缺陷,并提高运动线索的利用率。采用多流一致性融合模型优化三流分支,得到不同特征的最佳融合方案。同时通过循环训练策略平衡三重网络的权重,以避免网络过度拟合单流分支,从而有效地提高运动估计和定位的准确率。在Davis数据集上的实验结果表明,相比PCSA、SSAV、MGA等算法,该算法的鲁棒性更优,其maxF和S-Measure值分别达到0.893和0.912,MAE仅为0.021。     

Bioinspired yeast microcapsules loaded with self-assembled nanotherapies for targeted treatment of cardiovascular disease

1. Download high-res image (237KB)
2. Download full-size image

     

Application potential of bone marrow mesenchymal stem cell (BMSCs) based tissue-engineering for spinal cord defect repair in rat fetuses with spina bifida aperta

Spina bifida aperta are complex congenital malformations resulting from failure of fusion in the spinal neural tube during embryogenesis. Despite surgical repair of the defect, most patients who survive with spina bifida aperta have a multiple system handicap due to neuron deficiency of the defective spinal cord. Tissue engineering has emerged as a novel treatment for replacement of lost tissue. This study evaluated the prenatal surgical approach of transplanting a chitosan–gelatin scaffold seeded with bone marrow mesenchymal stem cells (BMSCs) in the healing the defective spinal cord of rat fetuses with retinoic acid induced spina bifida aperta. Scaffold characterisation revealed the porous structure, organic and amorphous content. This biomaterial promoted the adhesion, spreading and in vitro viability of the BMSCs. After transplantation of the scaffold combined with BMSCs, the defective region of spinal cord in rat fetuses with spina bifida aperta at E20 decreased obviously under stereomicroscopy, and the skin defect almost closed in many fetuses. The transplanted BMSCs in chitosan–gelatin scaffold survived, grew and expressed markers of neural stem cells and neurons in the defective spinal cord. In addition, the biomaterial presented high biocompatibility and slow biodegradation in vivo. In conclusion, prenatal transplantation of the scaffold combined with BMSCs could treat spinal cord defect in fetuses with spina bifida aperta by the regeneration of neurons and repairmen of defective region.