针对基于You Only Look Once v2算法的目标检测存在精度低及稳健性差的问题,提出一种车辆目标实时检测的You Only Look Once v2优化算法;该算法以You Only Look Once v2算法为基础,通过增加网络深度,增强特征提取能力,同时,通过添加残...针对基于You Only Look Once v2算法的目标检测存在精度低及稳健性差的问题,提出一种车辆目标实时检测的You Only Look Once v2优化算法;该算法以You Only Look Once v2算法为基础,通过增加网络深度,增强特征提取能力,同时,通过添加残差模块,解决网络深度增加带来的梯度消失或弥散问题;该方法将网络结构中低层特征与高层特征进行融合,提升对小目标车辆的检测精度。结果表明,通过在KITTI数据集上进行测试,优化后的算法在检测速度不变的情况下,提高了车辆目标检测精度,平均精度达到0.94,同时提升了小目标检测的准确性。展开更多
OBJECTIVE Mu-Xiang-You-Fang(MXYF)is a classic prescription of Hui medicine,composed of five herbs,which has been used to treat ischemic stroke for many years.However,the potential pharmacological mecha⁃nisms of MXYF r...OBJECTIVE Mu-Xiang-You-Fang(MXYF)is a classic prescription of Hui medicine,composed of five herbs,which has been used to treat ischemic stroke for many years.However,the potential pharmacological mecha⁃nisms of MXYF remain unclear.The present research is to investigate the neuroprotective effect of MXYF and its role in modulating autophagy via AMPK/mTOR signaling pathway in the PC12 oxygen-glucose deprivation and reperfusion(OGD/R)injury model.METHODS MXYF was extracted by supercritical CO2 fluid extraction apparatus.PC12 OGD/R injury model was established by oxygen-glucose deprivation for 2 h and reperfusion for 24 h.The effects of MXYF on the viability and cytotoxicity of PC12 cells were determined through cell counting kit(CCK-8)assay.Colorimetric method was performed to determine the LDH leakage rate.The calcium concentration was determined by chemical fluorescence method and the mitochondrial membrane potential was determined through flow cytometry.Monodansylcadaverine(MDC)staining was conducted to detect autophagosome formation.The expression of LC3,Beclin1,p62,p-AMPK,ULK1,p-mTOR and p-p70s6k proteins were determined by immunofluorescence and Western blotting analyses.RESULTS MXYF(1,2 and 4 mg·L^-1)could significantly increase the cell viability and mitochondrial membrane potential,while decreased the release of lactate dehydrogenase(LDH)and calcium concentration in PC12 cells.Mechanistic studies showed that MXYF reduced the LC3-II/LC3-I ratio and inhibited the expression of beclin1,p-AMPK and ULK1.In comparison,the expres⁃sion of p-mTOR,p-p70s6k and p62 were significantly enhanced.CONCLUSION MXYF inhibits autophagy after OGD/Rinduced PC12 cell injury through AMPK-mTOR pathway,thus MXYF might have therapeutic potential for treating the ischemic stroke.展开更多
针对果园现场苹果分级存在的计算资源受限和表面缺陷尺度差异大的问题,本研究构建基于机器视觉的改进YOLOv8苹果表面缺陷识别模型,在提高苹果表面缺陷检测效率的同时保证检测准确率。采用自搭建的机器视觉系统采集5500张苹果样本的表面...针对果园现场苹果分级存在的计算资源受限和表面缺陷尺度差异大的问题,本研究构建基于机器视觉的改进YOLOv8苹果表面缺陷识别模型,在提高苹果表面缺陷检测效率的同时保证检测准确率。采用自搭建的机器视觉系统采集5500张苹果样本的表面特征及缺陷图像,涵盖果柄、花萼的特征与黑点、腐烂、机械损伤、日灼、褐斑和裂纹6种常见表面缺陷以及1种环境杂物并完成特征标注。引入RepGhostNeXt和EffQAFPN算法结构,对YOLOv8(You Only Look Once version 8)检测模型的主干特征提取网络和特征金字塔进行改进。在此基础上,研究训练并比较了YOLOv8、YOLOv8n、YOLOv8+EffQAFPN、YOLOv8+Rep Ghost NeXt和YOLOv8+EffQAFPN+Rep Ghost NeXt5种模型,并重点对比模型在苹果表面瑕疵检测中的检测准确率和模型检测速度。研究结果表明,YOLOv8+EffQAFPN+RepGhostNeXt模型在综合检测性能上表现最佳,其整体识别准确率为94.9%,且保持了7.81帧/s的平均检测帧率。综上,该模型能够在计算资源有限的环境下高效完成苹果表面缺陷检测任务,为实现果园现场高效便捷的苹果分级提供技术支撑。展开更多
为了解决施工场景下安全帽佩戴检测时,由于人员密集、遮挡和复杂背景等原因造成的小目标漏检和错检的问题,提出一种基于YOLOv8n的双重注意力机制的跨层多尺度安全帽佩戴检测算法。首先,设计微小目标检测头,以提高模型对小目标的检测能力...为了解决施工场景下安全帽佩戴检测时,由于人员密集、遮挡和复杂背景等原因造成的小目标漏检和错检的问题,提出一种基于YOLOv8n的双重注意力机制的跨层多尺度安全帽佩戴检测算法。首先,设计微小目标检测头,以提高模型对小目标的检测能力;其次,在特征提取网络中嵌入双重注意力机制,从而更加关注复杂场景下目标信息的特征捕获;然后,将特征融合网络替换成重参数化泛化特征金字塔网络(RepGFPN)改进后的跨层多尺度特征融合结构S-GFPN(Selective layer Generalized Feature Pyramid Network),以实现小目标特征层信息和其他特征层的多尺度融合,并建立长期的依赖关系,从而抑制背景信息的干扰;最后,采用MPDIOU(Intersection Over Union with Minimum Point Distance)损失函数来解决尺度变化不敏感的问题。在公开数据集GDUT-HWD上的实验结果表明,改进后的模型比YOLOv8n的mAP@0.5提升了3.4个百分点,对蓝色、黄色、白色和红色安全帽的检测精度分别提升了2.0、1.1、4.6和9.1个百分点,在密集、遮挡、小目标、反光和黑暗这5类复杂场景下的可视化检测效果也优于YOLOv8n,为实际施工场景中安全帽佩戴检测提供了一种有效方法。展开更多
针对目前遥感图像小目标检测任务中易出现漏检和误检的问题,提出一种SCS-YOLO[SMCA+CSC+SIoU(shape-aware intersection over union loss)-you only look once]的遥感图像小目标检测算法。首先,针对遥感图像中目标小而聚集的问题,构建...针对目前遥感图像小目标检测任务中易出现漏检和误检的问题,提出一种SCS-YOLO[SMCA+CSC+SIoU(shape-aware intersection over union loss)-you only look once]的遥感图像小目标检测算法。首先,针对遥感图像中目标小而聚集的问题,构建空间多尺度卷积注意力(spatial multi-scale convolutional attention,SMCA),提升模型对空间和通道信息的特征提取能力;其次,针对深层网络传递时小目标语义信息容易丢失的问题,设计聚合亚像素卷积(concentrated sub-pixel convolution,CSC),采用多尺度聚合特征提取方法,增强了网络对语义信息的提取能力;最后,将SIoU损失函数替代原模型中的CIoU(complete intersection over union loss)损失函数,加快了网络的收敛速度。SCS-YOLO模型在RSOD和NWPU VHR-10数据集上,平均精确率的平均值(mAP)分别达到97%和90.9%,相较于原模型分别提升了2.2%和2.7%,可见该方法在遥感图像小目标检测任务中的有效性。展开更多
文摘针对基于You Only Look Once v2算法的目标检测存在精度低及稳健性差的问题,提出一种车辆目标实时检测的You Only Look Once v2优化算法;该算法以You Only Look Once v2算法为基础,通过增加网络深度,增强特征提取能力,同时,通过添加残差模块,解决网络深度增加带来的梯度消失或弥散问题;该方法将网络结构中低层特征与高层特征进行融合,提升对小目标车辆的检测精度。结果表明,通过在KITTI数据集上进行测试,优化后的算法在检测速度不变的情况下,提高了车辆目标检测精度,平均精度达到0.94,同时提升了小目标检测的准确性。
基金National Natural Science Foundation of China(8166070081260679)Ningxia College FirstClass Discipline Construction Project(Chinese Medicine)Funded Project(NXYLXK2017A06)
文摘OBJECTIVE Mu-Xiang-You-Fang(MXYF)is a classic prescription of Hui medicine,composed of five herbs,which has been used to treat ischemic stroke for many years.However,the potential pharmacological mecha⁃nisms of MXYF remain unclear.The present research is to investigate the neuroprotective effect of MXYF and its role in modulating autophagy via AMPK/mTOR signaling pathway in the PC12 oxygen-glucose deprivation and reperfusion(OGD/R)injury model.METHODS MXYF was extracted by supercritical CO2 fluid extraction apparatus.PC12 OGD/R injury model was established by oxygen-glucose deprivation for 2 h and reperfusion for 24 h.The effects of MXYF on the viability and cytotoxicity of PC12 cells were determined through cell counting kit(CCK-8)assay.Colorimetric method was performed to determine the LDH leakage rate.The calcium concentration was determined by chemical fluorescence method and the mitochondrial membrane potential was determined through flow cytometry.Monodansylcadaverine(MDC)staining was conducted to detect autophagosome formation.The expression of LC3,Beclin1,p62,p-AMPK,ULK1,p-mTOR and p-p70s6k proteins were determined by immunofluorescence and Western blotting analyses.RESULTS MXYF(1,2 and 4 mg·L^-1)could significantly increase the cell viability and mitochondrial membrane potential,while decreased the release of lactate dehydrogenase(LDH)and calcium concentration in PC12 cells.Mechanistic studies showed that MXYF reduced the LC3-II/LC3-I ratio and inhibited the expression of beclin1,p-AMPK and ULK1.In comparison,the expres⁃sion of p-mTOR,p-p70s6k and p62 were significantly enhanced.CONCLUSION MXYF inhibits autophagy after OGD/Rinduced PC12 cell injury through AMPK-mTOR pathway,thus MXYF might have therapeutic potential for treating the ischemic stroke.
文摘针对果园现场苹果分级存在的计算资源受限和表面缺陷尺度差异大的问题,本研究构建基于机器视觉的改进YOLOv8苹果表面缺陷识别模型,在提高苹果表面缺陷检测效率的同时保证检测准确率。采用自搭建的机器视觉系统采集5500张苹果样本的表面特征及缺陷图像,涵盖果柄、花萼的特征与黑点、腐烂、机械损伤、日灼、褐斑和裂纹6种常见表面缺陷以及1种环境杂物并完成特征标注。引入RepGhostNeXt和EffQAFPN算法结构,对YOLOv8(You Only Look Once version 8)检测模型的主干特征提取网络和特征金字塔进行改进。在此基础上,研究训练并比较了YOLOv8、YOLOv8n、YOLOv8+EffQAFPN、YOLOv8+Rep Ghost NeXt和YOLOv8+EffQAFPN+Rep Ghost NeXt5种模型,并重点对比模型在苹果表面瑕疵检测中的检测准确率和模型检测速度。研究结果表明,YOLOv8+EffQAFPN+RepGhostNeXt模型在综合检测性能上表现最佳,其整体识别准确率为94.9%,且保持了7.81帧/s的平均检测帧率。综上,该模型能够在计算资源有限的环境下高效完成苹果表面缺陷检测任务,为实现果园现场高效便捷的苹果分级提供技术支撑。
文摘为了解决施工场景下安全帽佩戴检测时,由于人员密集、遮挡和复杂背景等原因造成的小目标漏检和错检的问题,提出一种基于YOLOv8n的双重注意力机制的跨层多尺度安全帽佩戴检测算法。首先,设计微小目标检测头,以提高模型对小目标的检测能力;其次,在特征提取网络中嵌入双重注意力机制,从而更加关注复杂场景下目标信息的特征捕获;然后,将特征融合网络替换成重参数化泛化特征金字塔网络(RepGFPN)改进后的跨层多尺度特征融合结构S-GFPN(Selective layer Generalized Feature Pyramid Network),以实现小目标特征层信息和其他特征层的多尺度融合,并建立长期的依赖关系,从而抑制背景信息的干扰;最后,采用MPDIOU(Intersection Over Union with Minimum Point Distance)损失函数来解决尺度变化不敏感的问题。在公开数据集GDUT-HWD上的实验结果表明,改进后的模型比YOLOv8n的mAP@0.5提升了3.4个百分点,对蓝色、黄色、白色和红色安全帽的检测精度分别提升了2.0、1.1、4.6和9.1个百分点,在密集、遮挡、小目标、反光和黑暗这5类复杂场景下的可视化检测效果也优于YOLOv8n,为实际施工场景中安全帽佩戴检测提供了一种有效方法。
文摘针对目前遥感图像小目标检测任务中易出现漏检和误检的问题,提出一种SCS-YOLO[SMCA+CSC+SIoU(shape-aware intersection over union loss)-you only look once]的遥感图像小目标检测算法。首先,针对遥感图像中目标小而聚集的问题,构建空间多尺度卷积注意力(spatial multi-scale convolutional attention,SMCA),提升模型对空间和通道信息的特征提取能力;其次,针对深层网络传递时小目标语义信息容易丢失的问题,设计聚合亚像素卷积(concentrated sub-pixel convolution,CSC),采用多尺度聚合特征提取方法,增强了网络对语义信息的提取能力;最后,将SIoU损失函数替代原模型中的CIoU(complete intersection over union loss)损失函数,加快了网络的收敛速度。SCS-YOLO模型在RSOD和NWPU VHR-10数据集上,平均精确率的平均值(mAP)分别达到97%和90.9%,相较于原模型分别提升了2.2%和2.7%,可见该方法在遥感图像小目标检测任务中的有效性。
