Objective To evaluate the optic nerve impairment using MRI histogram texture analysis in the patients with optic neuritis.Methods The study included 60 patients with optic neuritis and 20 normal controls. The coronal ...Objective To evaluate the optic nerve impairment using MRI histogram texture analysis in the patients with optic neuritis.Methods The study included 60 patients with optic neuritis and 20 normal controls. The coronal T2 weighted imaging(T2 WI) with fat saturation and enhanced T1 weighted imaging(T1 WI) were performed to evaluate the optic nerve. MRI histogram texture features of the involved optic nerve were measured on the corresponding coronal T2 WI images. The normal optic nerve(NON) was measured in the posterior 1/3 parts of the optic nerve. Kruskal-Wallis one-way ANOVA was used to compare the difference of texture features and receiver operating characteristic(ROC) curve were performed to evaluate the diagnostic value of texture features for the optic nerve impairment among the affected optic nerve with enhancement(ONwEN), affected optic nerve without enhancement(ONwoEN), contralateral normal appearing optic nerve(NAON) and NON. Results The histogram texture Energy and Entropy presented significant differences for ONwEN vs. ONwoEN(both P = 0.000), ONwEN vs. NON(both P = 0.000) and NAON vs. NON(both P < 0.05). ROC analysis demonstrated that the area under the curve(AUC) of histogram texture Energy were 0.758, 0.795 and 0.701 for ONwEN vs. ONwoEN, ONwEN vs. NON and NAON vs. NON, AUC of Entropy were 0.758, 0.795 and 0.707 for ONwEN vs. ONwoEN, ONwEN vs. NON and NAON vs. NON.Conclusion The altered MRI histogram texture Energy and Entropy could be considered as a surrogate for MRI enhancement to evaluate the involved optic nerve and normal-appearing optic nerve in optic neuritis.展开更多
Free energy calculations may provide vital information for studying various chemical and biological processes.Quantum mechanical methods are required to accurately describe interaction energies,but their computations ...Free energy calculations may provide vital information for studying various chemical and biological processes.Quantum mechanical methods are required to accurately describe interaction energies,but their computations are often too demanding for conformational sampling.As a remedy,level correction schemes that allow calculating high level free energies based on conformations from lower level simulations have been developed.Here,we present a variation of a Monte Carlo(MC)resampling approach in relation to the weighted histogram analysis method(WHAM).We show that our scheme can generate free energy surfaces that can practically converge to the exact one with sufficient sampling,and that it treats cases with insufficient sampling in a more stable manner than the conventional WHAM-based level correction scheme.It can also provide a guide for checking the uncertainty of the levelcorrected surface and a well-defined criterion for deciding the extent of smoothing on the free energy surface for its visual improvement.We demonstrate these aspects by obtaining the free energy maps associated with the alanine dipeptide and proton transfer network of the KillerRed protein in explicit water,and exemplify that the MC resampled WHAM scheme can be a practical tool for producing free energy surfaces of realistic systems.展开更多
利用超声波驻波声场的声能量密度在空间呈周期分布的性质,结合纹影法,对空气介质中28 k Hz超声波形成的驻波声场进行成像,使得超声波驻波声场的某些特性可视化,进而更加精准测量其波长和声速值。用光线折射几何理论证明了光的不均匀性...利用超声波驻波声场的声能量密度在空间呈周期分布的性质,结合纹影法,对空气介质中28 k Hz超声波形成的驻波声场进行成像,使得超声波驻波声场的某些特性可视化,进而更加精准测量其波长和声速值。用光线折射几何理论证明了光的不均匀性所导致的光的折射和弯曲程度与x、y平面内各自折射率梯度成正比,并推导了光线的偏折角计算公式。为使图像更加形象可视,在MATLAB中先对纹影图像做灰度直方图均衡化,再进行灰度-颜色映射处理,得到了较好伪彩色纹影图,使实验数据测量更加便利。最后利用本文方法推导了声速测量公式,准确测量了速度值,并用数理方法分析了仪器系统的灵敏度。展开更多
多源影像匹配主要受到非线性强度差异、对比度差异及局部区域结构特征不显著等问题的干扰,而机载激光雷达(light detection and ranging,LiDAR)深度图与航空影像由于纹理特征之间的显著差异导致部分结构特征缺失更为严重,所造成的相位...多源影像匹配主要受到非线性强度差异、对比度差异及局部区域结构特征不显著等问题的干扰,而机载激光雷达(light detection and ranging,LiDAR)深度图与航空影像由于纹理特征之间的显著差异导致部分结构特征缺失更为严重,所造成的相位极值突变进一步增加了匹配难度。因此,提出一种基于相位均匀卷积描述子的方法来实现LiDAR深度图与航空影像之间的高效匹配。在影像特征匹配阶段,首先对相位一致性模型进行扩展,构造相位均匀能量卷积方程,求解得到均匀卷积序列与相位最大标签图,建立一种相位均匀能量卷积直方图(histogram of phase mean energy convolution,HPMEC);然后采用最近邻匹配算法完成初始匹配,并利用快速边缘化样本共识进行粗差剔除;最后基于线程池并行策略,通过划分重叠格网对影像进行分块匹配。将多组具有不同地物覆盖类型的LiDAR深度图与航空影像作为数据集,分别与位置尺度方向-尺度不变特征转换(position scale orientation-scale invariant feature transform,PSO-SIFT)、Log-Gabor直方图描述(Log-Gabor histogram descriptor,LGHD)、辐射变化强度特征转换(radiation-variation insensitive feature transform,RIFT)和绝对相位一致性梯度直方图(histogram of absolute phase consistency gradients,HAPCG)等方法进行对比实验。结果表明,在LiDAR深度图与航空影像匹配中,HPMEC方法性能明显优于其他4种方法,其平均运行时间是PSO-SIFT的13.3倍、LGHD的10.9倍、HAPCG的10.4倍和RIFT的7.0倍;平均正确匹配点数显著高于其他4种方法;均方根误差在1.9像素以内,略优于其他4种方法。HPMEC方法在LiDAR深度图与航空影像中能够实现高效、稳健匹配。展开更多
The classification of spatial characteristics and discharge modes of dielectric barrier discharge(DBD)are gaining increasing attention in industrial applications,especially in the field of surface treatment of materia...The classification of spatial characteristics and discharge modes of dielectric barrier discharge(DBD)are gaining increasing attention in industrial applications,especially in the field of surface treatment of materials.In this work,gray level histogram(GLH)and Fourier energy spectrum based on the digital image processing tech no logy are applied to investigate the spatial structure and discharge mode of mesh-plate DBD.The coefficient of variation(CV)is calculated to describe the uniformity of the discharge.The results show that the discharge mode of mesh-plate DBD changes from periodic discharge to filamentary discharge when the applied voltage increases from 11-15 kV.Moreover,a more regular spatial structure is obtained under lower applied voltages during the discharge process.It is also found that the apertures of mesh electrodes which are below 1 mm have smaller values of CV compared to plate electrodes,indicating more uniform discharge.Finally,polypropylene is treated by mesh-plate DBD for surface modification.The hydrophilicity is significantly improved as the water contact angle decreased by 64°,and the dyeing depth is also enhanced.展开更多
文摘Objective To evaluate the optic nerve impairment using MRI histogram texture analysis in the patients with optic neuritis.Methods The study included 60 patients with optic neuritis and 20 normal controls. The coronal T2 weighted imaging(T2 WI) with fat saturation and enhanced T1 weighted imaging(T1 WI) were performed to evaluate the optic nerve. MRI histogram texture features of the involved optic nerve were measured on the corresponding coronal T2 WI images. The normal optic nerve(NON) was measured in the posterior 1/3 parts of the optic nerve. Kruskal-Wallis one-way ANOVA was used to compare the difference of texture features and receiver operating characteristic(ROC) curve were performed to evaluate the diagnostic value of texture features for the optic nerve impairment among the affected optic nerve with enhancement(ONwEN), affected optic nerve without enhancement(ONwoEN), contralateral normal appearing optic nerve(NAON) and NON. Results The histogram texture Energy and Entropy presented significant differences for ONwEN vs. ONwoEN(both P = 0.000), ONwEN vs. NON(both P = 0.000) and NAON vs. NON(both P < 0.05). ROC analysis demonstrated that the area under the curve(AUC) of histogram texture Energy were 0.758, 0.795 and 0.701 for ONwEN vs. ONwoEN, ONwEN vs. NON and NAON vs. NON, AUC of Entropy were 0.758, 0.795 and 0.707 for ONwEN vs. ONwoEN, ONwEN vs. NON and NAON vs. NON.Conclusion The altered MRI histogram texture Energy and Entropy could be considered as a surrogate for MRI enhancement to evaluate the involved optic nerve and normal-appearing optic nerve in optic neuritis.
基金supported by the Mid-career Researcher Program(No.2017R1A2B3004946)through National Research Foundationfunded by Ministry of Science and ICT of Korea.
文摘Free energy calculations may provide vital information for studying various chemical and biological processes.Quantum mechanical methods are required to accurately describe interaction energies,but their computations are often too demanding for conformational sampling.As a remedy,level correction schemes that allow calculating high level free energies based on conformations from lower level simulations have been developed.Here,we present a variation of a Monte Carlo(MC)resampling approach in relation to the weighted histogram analysis method(WHAM).We show that our scheme can generate free energy surfaces that can practically converge to the exact one with sufficient sampling,and that it treats cases with insufficient sampling in a more stable manner than the conventional WHAM-based level correction scheme.It can also provide a guide for checking the uncertainty of the levelcorrected surface and a well-defined criterion for deciding the extent of smoothing on the free energy surface for its visual improvement.We demonstrate these aspects by obtaining the free energy maps associated with the alanine dipeptide and proton transfer network of the KillerRed protein in explicit water,and exemplify that the MC resampled WHAM scheme can be a practical tool for producing free energy surfaces of realistic systems.
文摘利用超声波驻波声场的声能量密度在空间呈周期分布的性质,结合纹影法,对空气介质中28 k Hz超声波形成的驻波声场进行成像,使得超声波驻波声场的某些特性可视化,进而更加精准测量其波长和声速值。用光线折射几何理论证明了光的不均匀性所导致的光的折射和弯曲程度与x、y平面内各自折射率梯度成正比,并推导了光线的偏折角计算公式。为使图像更加形象可视,在MATLAB中先对纹影图像做灰度直方图均衡化,再进行灰度-颜色映射处理,得到了较好伪彩色纹影图,使实验数据测量更加便利。最后利用本文方法推导了声速测量公式,准确测量了速度值,并用数理方法分析了仪器系统的灵敏度。
文摘多源影像匹配主要受到非线性强度差异、对比度差异及局部区域结构特征不显著等问题的干扰,而机载激光雷达(light detection and ranging,LiDAR)深度图与航空影像由于纹理特征之间的显著差异导致部分结构特征缺失更为严重,所造成的相位极值突变进一步增加了匹配难度。因此,提出一种基于相位均匀卷积描述子的方法来实现LiDAR深度图与航空影像之间的高效匹配。在影像特征匹配阶段,首先对相位一致性模型进行扩展,构造相位均匀能量卷积方程,求解得到均匀卷积序列与相位最大标签图,建立一种相位均匀能量卷积直方图(histogram of phase mean energy convolution,HPMEC);然后采用最近邻匹配算法完成初始匹配,并利用快速边缘化样本共识进行粗差剔除;最后基于线程池并行策略,通过划分重叠格网对影像进行分块匹配。将多组具有不同地物覆盖类型的LiDAR深度图与航空影像作为数据集,分别与位置尺度方向-尺度不变特征转换(position scale orientation-scale invariant feature transform,PSO-SIFT)、Log-Gabor直方图描述(Log-Gabor histogram descriptor,LGHD)、辐射变化强度特征转换(radiation-variation insensitive feature transform,RIFT)和绝对相位一致性梯度直方图(histogram of absolute phase consistency gradients,HAPCG)等方法进行对比实验。结果表明,在LiDAR深度图与航空影像匹配中,HPMEC方法性能明显优于其他4种方法,其平均运行时间是PSO-SIFT的13.3倍、LGHD的10.9倍、HAPCG的10.4倍和RIFT的7.0倍;平均正确匹配点数显著高于其他4种方法;均方根误差在1.9像素以内,略优于其他4种方法。HPMEC方法在LiDAR深度图与航空影像中能够实现高效、稳健匹配。
基金financial support from the Joint Funds of National Natural Science Foundation of China(No.U1462105)
文摘The classification of spatial characteristics and discharge modes of dielectric barrier discharge(DBD)are gaining increasing attention in industrial applications,especially in the field of surface treatment of materials.In this work,gray level histogram(GLH)and Fourier energy spectrum based on the digital image processing tech no logy are applied to investigate the spatial structure and discharge mode of mesh-plate DBD.The coefficient of variation(CV)is calculated to describe the uniformity of the discharge.The results show that the discharge mode of mesh-plate DBD changes from periodic discharge to filamentary discharge when the applied voltage increases from 11-15 kV.Moreover,a more regular spatial structure is obtained under lower applied voltages during the discharge process.It is also found that the apertures of mesh electrodes which are below 1 mm have smaller values of CV compared to plate electrodes,indicating more uniform discharge.Finally,polypropylene is treated by mesh-plate DBD for surface modification.The hydrophilicity is significantly improved as the water contact angle decreased by 64°,and the dyeing depth is also enhanced.
