Differences between revisions 1 and 12 (spanning 11 versions)
Revision 1 as of 2007-12-10 22:47:19
Size: 670
Comment:
Revision 12 as of 2009-03-24 02:11:25
Size: 1703
Comment:
Deletions are marked like this. Additions are marked like this.
Line 1: Line 1:
{{{
[someone@localhost]$ e2.py
= Transforming Images (Applying Rotation, Translation, Scale and Mirroring) =

What follows is a very simply demonstration of using the Transform object to transform EMData (image) objects. A very thorough explanation of using the Transform is presented in [[Eman2TransformInPython|Using the EMAN2 Transform class]] and you are encouraged to at least be aware of that page's existence.

== Using EMData.transform ==

The first example here is a '''3D''' example.

{{{#!python
[someone@localhost]# e2.py
Line 8: Line 16:
In [3]: e = EMData()
Line 10: Line 17:
In [4]: e.set_size(32,32,32) In [3]: e = test_image_3d()
Line 12: Line 19:
In [5]: e.process_inplace('testimage.axes') In [4]: t = Transform({"type":"eman","az":45,"alt":90})
Line 14: Line 21:
In [6]: t = Transform3D(90,30,90) In [5]: t.set_trans(1,-2,5)
Line 16: Line 23:
In [7]: t.set_posttrans(1,-2,5) In [6]: t.set_scale(1)
Line 18: Line 25:
In [8]: t.set_pretrans(2,10,-20) In [7]: e.transform(t)
Line 20: Line 27:
In [9]: e.rotate_translate(t)

In [9
]: display(e)
In [8]: display(e)
Line 26: Line 31:
Notes: This second example is a '''2D''' example.
Line 28: Line 33:
1. The Transform3D is initialized on line 6 with the arguments 90 (azimuth), 30 (altitude), and 90 (phi). This is ZXZ' rotation. You can do rotations using SPIDER, MRC, IMAGIC and many other conventions with the Transform3D object, but this is not demonstrated here. {{{#!python

In [9]: e = test_image() # This is a 2D test image

In [10]: t = Transform({"type":"2d","alpha":45}) # Note 2D terminology

In [11]: t.set_trans(3,-2) # Note 2D translation

In [12]: t.set_mirror(True)

In [13]: e.transform(t)

In [14]: display(e)

}}}


== Using the Processor Framework ==

You can use the processor framework to process an EMData object in and out of place. In the latter case it is more efficient to use the processing framework as opposed to make a copy followed by EMData.transform.

{{{#!python

In [15]: t = Transform(....) # Construct a transform as above

In [16]: out_of_place = e.process("math.transform",{"transform":t}) # Out of place

In [17]: e.process_inplace("math.transform",{"transform":t}) # Inplace

In [18]: e.transform(t) # exactly the same as line above

}}}

Transforming Images (Applying Rotation, Translation, Scale and Mirroring)

What follows is a very simply demonstration of using the Transform object to transform EMData (image) objects. A very thorough explanation of using the Transform is presented in Using the EMAN2 Transform class and you are encouraged to at least be aware of that page's existence.

Using EMData.transform

The first example here is a 3D example.

   1 [someone@localhost]# e2.py
   2 
   3 Welcome to EMAN2
   4 Prompt provided by IPython
   5 Enter '?' for ipython help
   6 
   7 
   8 In [3]:  e = test_image_3d()
   9 
  10 In [4]:  t = Transform({"type":"eman","az":45,"alt":90})
  11 
  12 In [5]:  t.set_trans(1,-2,5)
  13 
  14 In [6]:  t.set_scale(1)
  15 
  16 In [7]:  e.transform(t)
  17 
  18 In [8]:  display(e)

This second example is a 2D example.

   1 In [9]:  e = test_image() # This is a 2D test image
   2 
   3 In [10]:  t = Transform({"type":"2d","alpha":45}) # Note 2D terminology
   4 
   5 In [11]:  t.set_trans(3,-2) # Note 2D translation
   6 
   7 In [12]:  t.set_mirror(True)
   8 
   9 In [13]:  e.transform(t)
  10 
  11 In [14]:  display(e)

Using the Processor Framework

You can use the processor framework to process an EMData object in and out of place. In the latter case it is more efficient to use the processing framework as opposed to make a copy followed by EMData.transform.

   1 In [15]:  t = Transform(....) # Construct a transform as above
   2 
   3 In [16]:  out_of_place = e.process("math.transform",{"transform":t}) # Out of place
   4 
   5 In [17]:  e.process_inplace("math.transform",{"transform":t}) # Inplace
   6 
   7 In [18]:  e.transform(t) # exactly the same as line above

EMAN2/Tutorials/RotateTranslate (last edited 2015-05-04 18:17:02 by StephenMurray)