由于外在因素导致出土的文物模型呈现破碎状态,对其进行虚拟修复对考古学具有重要意义。现有孔洞修补方法大多仅针对三维模型结构进行补全,并且在三维结构修复后缺少表面的颜色纹理信息。本文基于文物三维空间结构和纹理信息,提出一种...由于外在因素导致出土的文物模型呈现破碎状态,对其进行虚拟修复对考古学具有重要意义。现有孔洞修补方法大多仅针对三维模型结构进行补全,并且在三维结构修复后缺少表面的颜色纹理信息。本文基于文物三维空间结构和纹理信息,提出一种三维文物孔洞修复方法。首先,为了解决三维文物结构修复问题,利用基于径向基函数的算法填补三维文物网格模型的孔洞,通过拟合的曲面方程调整孔洞补丁顶点,使其与原有模型更好地融合。其次,为补全文物表面颜色纹理信息,使孔洞补丁与原有模型表面纹理自然过渡,将三维问题转化为二维图像修复问题,以EdgeConnect为框架,通过添加精细化网络生成更高分辨率的结果。最后,使用Mudbox软件映射二维图像到三维模型表面,融合结构和纹理修复的结果。本文改进后的二维修复网络在评价指标PSNR(Peak Signal to Noise Ratio)、SSIM(Structural Similarity)和MAE(Mean Absolute Deviation)的性能分别提高了0.54%、0.217%和6.52%,该方法能够有效地恢复兵马俑三维模型网格结构和表面纹理信息。展开更多
The paper presented here have introduced the principles and characteristics of several QTL(Quantitative trait locus) analysis methods in plant, including the single-marker QTL analysis, the interval mapping and the co...The paper presented here have introduced the principles and characteristics of several QTL(Quantitative trait locus) analysis methods in plant, including the single-marker QTL analysis, the interval mapping and the composite interval mapping. The single-marker QTL analysis is based on the sub-groups of marker genotypes to do the significance analysis of average values and the one-way analysis of variance. The interval mapping is used to estimate the probability of having QTL at any position in the interval between any two adjacent markers. The composite interval mapping has been developed on the basis of the interval mapping with use of multiple regression model in the ma- ximum likelihood estimate, and thus may give high accuracy and power in mapping QTLs due to the control over large number of background genetic variation in studying population. Especially in mapping QTLs with epistatic effects and QE interaction, some extensive ways of the composite interval mapping can produce more reliable展开更多
文摘由于外在因素导致出土的文物模型呈现破碎状态,对其进行虚拟修复对考古学具有重要意义。现有孔洞修补方法大多仅针对三维模型结构进行补全,并且在三维结构修复后缺少表面的颜色纹理信息。本文基于文物三维空间结构和纹理信息,提出一种三维文物孔洞修复方法。首先,为了解决三维文物结构修复问题,利用基于径向基函数的算法填补三维文物网格模型的孔洞,通过拟合的曲面方程调整孔洞补丁顶点,使其与原有模型更好地融合。其次,为补全文物表面颜色纹理信息,使孔洞补丁与原有模型表面纹理自然过渡,将三维问题转化为二维图像修复问题,以EdgeConnect为框架,通过添加精细化网络生成更高分辨率的结果。最后,使用Mudbox软件映射二维图像到三维模型表面,融合结构和纹理修复的结果。本文改进后的二维修复网络在评价指标PSNR(Peak Signal to Noise Ratio)、SSIM(Structural Similarity)和MAE(Mean Absolute Deviation)的性能分别提高了0.54%、0.217%和6.52%,该方法能够有效地恢复兵马俑三维模型网格结构和表面纹理信息。
文摘The paper presented here have introduced the principles and characteristics of several QTL(Quantitative trait locus) analysis methods in plant, including the single-marker QTL analysis, the interval mapping and the composite interval mapping. The single-marker QTL analysis is based on the sub-groups of marker genotypes to do the significance analysis of average values and the one-way analysis of variance. The interval mapping is used to estimate the probability of having QTL at any position in the interval between any two adjacent markers. The composite interval mapping has been developed on the basis of the interval mapping with use of multiple regression model in the ma- ximum likelihood estimate, and thus may give high accuracy and power in mapping QTLs due to the control over large number of background genetic variation in studying population. Especially in mapping QTLs with epistatic effects and QE interaction, some extensive ways of the composite interval mapping can produce more reliable
