将Java中的3D仿射变换构建导入R

Question

我有一个与仿射变换有关的问题（在<affine>标记之间）。我正在使用full xml file here函数从用Java创建的BigDataViewer Fiji插件创建的xml文件（AffineTransform3D）中提取以下两个仿射转换：

<ViewRegistrations>
<ViewRegistration timepoint="0" setup="0">
  <ViewTransform type="affine">
    <Name>Fast 3d geometric hashing (rotation invariant), AffineModel3D on [beads (c=0)]</Name>
    <affine>1.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 1.0 0.0</affine>
  </ViewTransform>
  <ViewTransform type="affine">
    <Name>calibration</Name>
    <affine>1.9998662296836334 0.0 0.0 0.0 0.0 1.9998662296836334 0.0 0.0 0.0 0.0 1.9998662296836334 0.0</affine>
  </ViewTransform>
</ViewRegistration>

我想使用R包{RNiftyReg}中的buildAffine()函数在R中导入两个仿射变换，然后使用{RNiftyReg}中的composeTransforms()计算它们的成分。

buildAffine(translation = c(0, 0, 0), scales = c(1, 1, 1), skews = c(0, 0,0),
  angles = c(0, 0, 0), source = NULL, target = NULL,
  anchor = c("none", "origin", "centre", "center"))

我的问题：
上面的仿射变换存储在12个索引的向量中。 buildAffine()要求将平移，比例，偏斜和角度的值作为输入参数。
我想知道哪个值对应什么。

Answer 1

我主要是R用户，但是这里有：Java调用中的变量名称是：

 dat <- scan (text="double xx, double yx, double zx, double tx, double xy, double yy, double zy, double ty, double xz, double yz, double zz, double tz", what="")
dat <- dat[dat != "double"]
matrix(dat,4)
      [,1]  [,2]  [,3] 
[1,] "xx," "xy," "xz,"
[2,] "yx," "yy," "yz,"
[3,] "zx," "zy," "zz,"
[4,] "tx," "ty," "tz" 

本页记录了仿射变换的编码方式，唯一不同之处在于转置了4x3矩阵：https://o7planning.org/en/11157/javafx-transformations-tutorial

 t( matrix(dat,4) )
     [,1]  [,2]  [,3]  [,4] 
[1,] "xx," "yx," "zx," "tx,"
[2,] "xy," "yy," "zy," "ty,"
[3,] "xz," "yz," "zz," "tz" 

因此，将矩阵应用于c（x，y，z，1）的向量以获得以下输出：