In recent years,the concept of digital human has attracted widespread attention from all walks of life,and the modelling of high-fidelity human bodies,heads,and hands has been intensively studied.This paper focuses on...In recent years,the concept of digital human has attracted widespread attention from all walks of life,and the modelling of high-fidelity human bodies,heads,and hands has been intensively studied.This paper focuses on head modelling and proposes a generic head parametric model based on neural radiance fields.Specifically,we first use face recognition networks and 3D facial expression database FaceWarehouse to parameterize identity and expression semantics,respectively,and use both as conditional inputs to build a neural radiance field for the human head,thereby improving the head model’s representation ability while ensuring editing capabilities for the identity and expression of the rendered results;then,through a combination of volume rendering and neural rendering,the 3D representation of the head is rapidly rendered into the 2D plane,producing a high-fidelity image of the human head.Thanks to the well-designed loss functions and good implicit representation of the neural radiance field,our model can not only edit the identity and expression independently,but also freely modify the virtual camera position of the rendering results.It has excellent multi-view consistency,and has many applications in novel view synthesis,pose driving and more.展开更多
目的探讨小视野扩散峰度成像(reduced field-of-view diffusion kurtosis imaging,rFOV-DKI)技术在鉴别子宫内膜样腺癌组织学分级中的潜力。材料与方法本研究共纳入48例经病理证实的子宫内膜样腺癌患者。依据国际妇产科联盟(The Interna...目的探讨小视野扩散峰度成像(reduced field-of-view diffusion kurtosis imaging,rFOV-DKI)技术在鉴别子宫内膜样腺癌组织学分级中的潜力。材料与方法本研究共纳入48例经病理证实的子宫内膜样腺癌患者。依据国际妇产科联盟(The International Federation of Gynecology and Obstetrics,FIGO)分级法,受试者分为低级别组(G1、G1~2和G2,n=30)和高级别组(G3,n=18)。所有受试者均于3.0 T MRI扫描仪下行常规盆腔平扫加增强及rFOV-DKI序列扫描。参照常规矢状位T2加权图像在rFOV-DKI序列图像手动勾画感兴趣区(region of interest,ROI)。计算ROI的扩散峰度成像(diffusion kurtosis imaging,DKI)衍生参数,包括平均扩散率(mean diffusivity,MD)、轴向扩散率(axial diffusivity,Da)、径向扩散率(radial diffusivity,Dr)、平均峰度(mean kurtosis,MK)、轴向峰度(axial kurtosis,Ka)和径向峰度(radial kurtosis,Kr)。比较分析各rFOV-DKI参数在低级别组和高级别组间的差异,同时采用受试者工作特征(receiver operating characteristic,ROC)曲线方法评估每个参数的诊断性能。采用DeLong方法对比各参数ROC曲线下面积(area under the curve,AUC)的差异。结果低级别组MD、Da和Dr的平均值[(0.93±0.08)μm^(2)/ms、(1.14±0.10)μm^(2)/ms、(0.83±0.08)μm^(2)/ms]高于高级别组的平均值[(0.80±0.08)μm^(2)/ms、(1.05±0.07)μm^(2)/ms、(0.74±0.06)μm^(2)/ms;P<0.05],而MK、Ka和Kr的平均值(1.15±0.10、1.36±0.10、0.97±0.13)则低于高级别组(1.33±0.11、1.64±0.11、1.08±0.09)(P<0.05)。Ka值在区分低级别组和高别级组时具有最高的诊断准确性,ROC曲线下面积(area under the curve,AUC)为0.98(95%CI:0.89~1.00),其次是MK[AUC=0.90(95%CI:0.78~0.97)]和MD[AUC=0.88(95%CI:0.76~0.96)]。MK与Ka和MD的AUC间差异均没有统计学意义(Z=1.81和0.53,P=0.07和0.59),而Ka和MD的AUC间差异具有统计学意义(Z=2.40,P=0.02)。在所有DKI衍生参数中,Ka在区分低级别组和高级别组方面表现最好,截断值为1.46,敏感度和特异度分别为100%和90%。结论基于非高斯扩散加权模型的rFOV-DKI可作为区分子宫内膜样腺癌组织学分级的潜在影像学工具,用于子宫内膜样腺癌的无创术前分级。展开更多
文摘In recent years,the concept of digital human has attracted widespread attention from all walks of life,and the modelling of high-fidelity human bodies,heads,and hands has been intensively studied.This paper focuses on head modelling and proposes a generic head parametric model based on neural radiance fields.Specifically,we first use face recognition networks and 3D facial expression database FaceWarehouse to parameterize identity and expression semantics,respectively,and use both as conditional inputs to build a neural radiance field for the human head,thereby improving the head model’s representation ability while ensuring editing capabilities for the identity and expression of the rendered results;then,through a combination of volume rendering and neural rendering,the 3D representation of the head is rapidly rendered into the 2D plane,producing a high-fidelity image of the human head.Thanks to the well-designed loss functions and good implicit representation of the neural radiance field,our model can not only edit the identity and expression independently,but also freely modify the virtual camera position of the rendering results.It has excellent multi-view consistency,and has many applications in novel view synthesis,pose driving and more.
文摘目的探讨小视野扩散峰度成像(reduced field-of-view diffusion kurtosis imaging,rFOV-DKI)技术在鉴别子宫内膜样腺癌组织学分级中的潜力。材料与方法本研究共纳入48例经病理证实的子宫内膜样腺癌患者。依据国际妇产科联盟(The International Federation of Gynecology and Obstetrics,FIGO)分级法,受试者分为低级别组(G1、G1~2和G2,n=30)和高级别组(G3,n=18)。所有受试者均于3.0 T MRI扫描仪下行常规盆腔平扫加增强及rFOV-DKI序列扫描。参照常规矢状位T2加权图像在rFOV-DKI序列图像手动勾画感兴趣区(region of interest,ROI)。计算ROI的扩散峰度成像(diffusion kurtosis imaging,DKI)衍生参数,包括平均扩散率(mean diffusivity,MD)、轴向扩散率(axial diffusivity,Da)、径向扩散率(radial diffusivity,Dr)、平均峰度(mean kurtosis,MK)、轴向峰度(axial kurtosis,Ka)和径向峰度(radial kurtosis,Kr)。比较分析各rFOV-DKI参数在低级别组和高级别组间的差异,同时采用受试者工作特征(receiver operating characteristic,ROC)曲线方法评估每个参数的诊断性能。采用DeLong方法对比各参数ROC曲线下面积(area under the curve,AUC)的差异。结果低级别组MD、Da和Dr的平均值[(0.93±0.08)μm^(2)/ms、(1.14±0.10)μm^(2)/ms、(0.83±0.08)μm^(2)/ms]高于高级别组的平均值[(0.80±0.08)μm^(2)/ms、(1.05±0.07)μm^(2)/ms、(0.74±0.06)μm^(2)/ms;P<0.05],而MK、Ka和Kr的平均值(1.15±0.10、1.36±0.10、0.97±0.13)则低于高级别组(1.33±0.11、1.64±0.11、1.08±0.09)(P<0.05)。Ka值在区分低级别组和高别级组时具有最高的诊断准确性,ROC曲线下面积(area under the curve,AUC)为0.98(95%CI:0.89~1.00),其次是MK[AUC=0.90(95%CI:0.78~0.97)]和MD[AUC=0.88(95%CI:0.76~0.96)]。MK与Ka和MD的AUC间差异均没有统计学意义(Z=1.81和0.53,P=0.07和0.59),而Ka和MD的AUC间差异具有统计学意义(Z=2.40,P=0.02)。在所有DKI衍生参数中,Ka在区分低级别组和高级别组方面表现最好,截断值为1.46,敏感度和特异度分别为100%和90%。结论基于非高斯扩散加权模型的rFOV-DKI可作为区分子宫内膜样腺癌组织学分级的潜在影像学工具,用于子宫内膜样腺癌的无创术前分级。
