Differences between revisions 1 and 2
Revision 1 as of 2007-12-14 21:29:31
Size: 760
Comment:
Revision 2 as of 2007-12-14 21:47:49
Size: 760
Comment:
Deletions are marked like this. Additions are marked like this.
Line 3: Line 3:
Answer (DaveWoolford) - I had a look at the source code and it looks like '''refine''' uses the shrink parameter to run '''classesbymra''' - there it is used to mean shrink the particles image prior to classification. Hence it speeds things up by disiregarding high frequency information. However, when the classes are finally generated the original images are used, and this involves involves an iterative alignment, so in a sense the high frequency information can still be resolved. So conceptually this flag speeds things up at at no penalty, but my guess is that it has not been thoroughly tested, and is not generally used. Answer (DaveWoolford) - I had a look at the source code and it looks like '''refine''' uses the shrink parameter to run '''classesbymra''' - there it is used to mean shrink the particle images prior to classification. Hence it speeds things up by disiregarding high frequency information. However, when the classes are finally generated the original images are used, and this involves involves an iterative alignment, so in a sense the high frequency information can still be resolved. So conceptually this flag speeds things up at at no penalty, but my guess is that it has not been thoroughly tested, and is not generally used.

What does the "shrink=n" do in Refine commmand? It doesn't look like changing the size of subfiles like proc2d shrink=n does?

Answer (DaveWoolford) - I had a look at the source code and it looks like refine uses the shrink parameter to run classesbymra - there it is used to mean shrink the particle images prior to classification. Hence it speeds things up by disiregarding high frequency information. However, when the classes are finally generated the original images are used, and this involves involves an iterative alignment, so in a sense the high frequency information can still be resolved. So conceptually this flag speeds things up at at no penalty, but my guess is that it has not been thoroughly tested, and is not generally used.

FAQ_EMAN_USING_refine_shrink (last edited 2008-11-26 04:42:29 by localhost)