In this paper,a data consistency condition(DCC)in Radon domain is re-derived from the perspective of Lie transformation group and compared with other two derivations.By carefully observing the proof procedure on the b...In this paper,a data consistency condition(DCC)in Radon domain is re-derived from the perspective of Lie transformation group and compared with other two derivations.By carefully observing the proof procedure on the basis of Lie transformation group,we may have a deeper insight on the essence of this DCC.Moreover,it may help determine whether there are the corresponding counterparts in both image and Radon domains.Following this line of thought,more data consistency conditions will possibly be explored in the future.This is definitely critical as data consistency conditions would be applied to the construction of algorithms for next-generation CT imaging.展开更多
针对面部点云配准中头发颜色与面部肤色存在差异,导致配准效率降低的问题,提出一种改进的相干点漂移(Coherent Point Drift,CPD)算法。该算法先提取面部点云的距离和颜色信息,将颜色信息转化为亮度信息,再仿照双侧滤波算法计算对应的亮...针对面部点云配准中头发颜色与面部肤色存在差异,导致配准效率降低的问题,提出一种改进的相干点漂移(Coherent Point Drift,CPD)算法。该算法先提取面部点云的距离和颜色信息,将颜色信息转化为亮度信息,再仿照双侧滤波算法计算对应的亮度权值系数。将亮度权值系数与距离权值系数相融合,得到新的加权系数,有效降低头发区域点的权重,以提高算法的配准精度与效率。最后,将改进的CPD算法应用于面部点云配准。实验结果表明,该算法能够将两点云的距离标准差平均降低66.01%,运算时间平均缩短18.57%,显著提高了面部点云配准效果。展开更多
基金supported in partby Shaanxi Provincial Natural Science Foundation of China(No.2023-JC-YB-521)by National Natural Science Foundation of China(NSFC)(No.12475313)+1 种基金Excellent Youth Fund of Shandong Natural Science Foundation(ZR2024YQ066)by Xi'an Science and Technology Program of China(No.23GXFW0065).
文摘In this paper,a data consistency condition(DCC)in Radon domain is re-derived from the perspective of Lie transformation group and compared with other two derivations.By carefully observing the proof procedure on the basis of Lie transformation group,we may have a deeper insight on the essence of this DCC.Moreover,it may help determine whether there are the corresponding counterparts in both image and Radon domains.Following this line of thought,more data consistency conditions will possibly be explored in the future.This is definitely critical as data consistency conditions would be applied to the construction of algorithms for next-generation CT imaging.
文摘针对面部点云配准中头发颜色与面部肤色存在差异,导致配准效率降低的问题,提出一种改进的相干点漂移(Coherent Point Drift,CPD)算法。该算法先提取面部点云的距离和颜色信息,将颜色信息转化为亮度信息,再仿照双侧滤波算法计算对应的亮度权值系数。将亮度权值系数与距离权值系数相融合,得到新的加权系数,有效降低头发区域点的权重,以提高算法的配准精度与效率。最后,将改进的CPD算法应用于面部点云配准。实验结果表明,该算法能够将两点云的距离标准差平均降低66.01%,运算时间平均缩短18.57%,显著提高了面部点云配准效果。
