Main Page | Modules | Namespace List | Class Hierarchy | Alphabetical List | Class List | Directories | File List | Namespace Members | Class Members | File Members

EMAN::RefineAligner Class Reference

refine alignment. More...

#include <aligner.h>

Inheritance diagram for EMAN::RefineAligner:

Inheritance graph
[legend]
Collaboration diagram for EMAN::RefineAligner:

Collaboration graph
[legend]
List of all members.

Public Member Functions

virtual EMDataalign (EMData *this_img, EMData *to_img, const string &cmp_name="dot", const Dict &cmp_params=Dict()) const
 To align 'this_img' with another image passed in through its parameters.
virtual EMDataalign (EMData *this_img, EMData *to_img) const
virtual string get_name () const
 Get the Aligner's name.
virtual string get_desc () const
virtual TypeDict get_param_types () const

Static Public Member Functions

AlignerNEW ()

Static Public Attributes

const string NAME = "refine"

Detailed Description

refine alignment.

Refines a preliminary 2D alignment using a simplex algorithm. Subpixel precision.

Parameters:
xform.align2d The Transform storing the starting guess. If unspecified the identity matrix is used
stepx The x increment used to create the starting simplex. Default is 1
stepy The y increment used to create the starting simplex. Default is 1
stepaz The rotational increment used to create the starting simplex. Default is 5
precision The precision which, if achieved, can stop the iterative refinement before reaching the maximum iterations. Default is 0.04
maxiter The maximum number of iterations. default=28
maxshift Maximum translation in pixels in any direction. If the solution yields a shift beyond this value in any direction, then the refinement is judged a failure and the original alignment is used as the solution
stepscale If set to any non-zero value, scale will be included in the alignment, and this will be the initial step. Images should be edgenormalized. If the scale goes beyond +-30% alignment will fail
verbose This will cause debugging information to be printed on the screen for the iterative refinement. Larger numbers -> more info. default=0

Definition at line 1221 of file aligner.h.


Member Function Documentation

virtual EMData* EMAN::RefineAligner::align EMData this_img,
EMData to_img
const [inline, virtual]
 

Implements EMAN::Aligner.

Definition at line 1227 of file aligner.h.

References align().

01228                 {
01229                         return align(this_img, to_img, "sqeuclidean", Dict());
01230                 }

EMData * RefineAligner::align EMData this_img,
EMData to_img,
const string &  cmp_name = "dot",
const Dict cmp_params = Dict()
const [virtual]
 

To align 'this_img' with another image passed in through its parameters.

The alignment uses a user-given comparison method to compare the two images. If none is given, a default one is used.

Parameters:
this_img The image to be compared.
to_img 'this_img" is aligned with 'to_img'.
cmp_name The comparison method to compare the two images.
cmp_params The parameter dictionary for comparison method.
Returns:
The aligned image.

Implements EMAN::Aligner.

Definition at line 1683 of file aligner.cpp.

References EMAN::EMData::get_attr(), EMAN::Transform::get_params(), EMAN::EMData::get_xsize(), EMAN::Dict::has_key(), EMAN::Cmp::params, EMAN::EMData::process(), EMAN::EMData::set_attr(), EMAN::Dict::set_default(), status, t, and x.

01685 {
01686 
01687         if (!to) {
01688                 return 0;
01689         }
01690 
01691         EMData *result;
01692         int mode = params.set_default("mode", 0);
01693         float saz = 0.0;
01694         float sdx = 0.0;
01695         float sdy = 0.0;
01696         float sscale = 1.0;
01697         bool mirror = false;
01698         Transform* t;
01699         if (params.has_key("xform.align2d") ) {
01700                 t = params["xform.align2d"];
01701                 Dict params = t->get_params("2d");
01702                 saz = params["alpha"];
01703                 sdx = params["tx"];
01704                 sdy = params["ty"];
01705                 mirror = params["mirror"];
01706                 sscale = params["scale"];
01707         } else {
01708                 t = new Transform(); // is the identity
01709         }
01710 
01711         // We do this to prevent the GSL routine from crashing on an invalid alignment
01712         if ((float)(this_img->get_attr("sigma"))==0.0 || (float)(to->get_attr("sigma"))==0.0) {
01713                 result = this_img->process("xform",Dict("transform",t));
01714                 result->set_attr("xform.align2d",t);
01715                 delete t;
01716                 return result;
01717         }
01718         
01719         float stepx = params.set_default("stepx",1.0f);
01720         float stepy = params.set_default("stepy",1.0f);
01721         // Default step is 5 degree - note in EMAN1 it was 0.1 radians
01722         float stepaz = params.set_default("stepaz",5.0f);
01723         float stepscale = params.set_default("stepscale",0.0f);
01724 
01725         int np = 3;
01726         if (stepscale!=0.0) np++;
01727         Dict gsl_params;
01728         gsl_params["this"] = this_img;
01729         gsl_params["with"] = to;
01730         gsl_params["snr"]  = params["snr"];
01731         gsl_params["mirror"] = mirror;
01732         if (params.has_key("mask")) gsl_params["mask"]=params["mask"];
01733         
01734         const gsl_multimin_fminimizer_type *T = gsl_multimin_fminimizer_nmsimplex;
01735         gsl_vector *ss = gsl_vector_alloc(np);
01736 
01737 
01738         gsl_vector_set(ss, 0, stepx);
01739         gsl_vector_set(ss, 1, stepy);
01740         gsl_vector_set(ss, 2, stepaz);
01741         if (stepscale!=0.0) gsl_vector_set(ss,3,stepscale);
01742         
01743         gsl_vector *x = gsl_vector_alloc(np);
01744         gsl_vector_set(x, 0, sdx);
01745         gsl_vector_set(x, 1, sdy);
01746         gsl_vector_set(x, 2, saz);
01747         if (stepscale!=0.0) gsl_vector_set(x,3,1.0);
01748         
01749         Cmp *c = 0;
01750 
01751         gsl_multimin_function minex_func;
01752         if (mode == 2) {
01753                 minex_func.f = &refalifnfast;
01754         }
01755         else {
01756                 c = Factory < Cmp >::get(cmp_name, cmp_params);
01757                 gsl_params["cmp"] = (void *) c;
01758                 minex_func.f = &refalifn;
01759         }
01760 
01761         minex_func.n = np;
01762         minex_func.params = (void *) &gsl_params;
01763 
01764         gsl_multimin_fminimizer *s = gsl_multimin_fminimizer_alloc(T, np);
01765         gsl_multimin_fminimizer_set(s, &minex_func, x, ss);
01766 
01767         int rval = GSL_CONTINUE;
01768         int status = GSL_SUCCESS;
01769         int iter = 1;
01770 
01771         float precision = params.set_default("precision",0.04f);
01772         int maxiter = params.set_default("maxiter",28);
01773 
01774 //      printf("Refine sx=%1.2f sy=%1.2f sa=%1.2f prec=%1.4f maxit=%d\n",stepx,stepy,stepaz,precision,maxiter);
01775 //      printf("%1.2f %1.2f %1.1f  ->",(float)gsl_vector_get(s->x, 0),(float)gsl_vector_get(s->x, 1),(float)gsl_vector_get(s->x, 2));
01776 
01777         while (rval == GSL_CONTINUE && iter < maxiter) {
01778                 iter++;
01779                 status = gsl_multimin_fminimizer_iterate(s);
01780                 if (status) {
01781                         break;
01782                 }
01783                 rval = gsl_multimin_test_size(gsl_multimin_fminimizer_size(s), precision);
01784         }
01785 
01786         int maxshift = params.set_default("maxshift",-1);
01787 
01788         if (maxshift <= 0) {
01789                 maxshift = this_img->get_xsize() / 4;
01790         }
01791         float fmaxshift = static_cast<float>(maxshift);
01792         if ( fmaxshift >= fabs((float)gsl_vector_get(s->x, 0)) && fmaxshift >= fabs((float)gsl_vector_get(s->x, 1)) && (stepscale==0 || (((float)gsl_vector_get(s->x, 3))<1.3 && ((float)gsl_vector_get(s->x, 3))<0.7))  )
01793         {
01794 //              printf(" Refine good %1.2f %1.2f %1.1f\n",(float)gsl_vector_get(s->x, 0),(float)gsl_vector_get(s->x, 1),(float)gsl_vector_get(s->x, 2));
01795                 Transform  tsoln(Dict("type","2d","alpha",(float)gsl_vector_get(s->x, 2)));
01796                 tsoln.set_mirror(mirror);
01797                 tsoln.set_trans((float)gsl_vector_get(s->x, 0),(float)gsl_vector_get(s->x, 1));
01798                 if (stepscale!=0.0) tsoln.set_scale((float)gsl_vector_get(s->x, 3));
01799                 result = this_img->process("xform",Dict("transform",&tsoln));
01800                 result->set_attr("xform.align2d",&tsoln);
01801         } else { // The refine aligner failed - this shift went beyond the max shift
01802 //              printf(" Refine Failed %1.2f %1.2f %1.1f\n",(float)gsl_vector_get(s->x, 0),(float)gsl_vector_get(s->x, 1),(float)gsl_vector_get(s->x, 2));
01803                 result = this_img->process("xform",Dict("transform",t));
01804                 result->set_attr("xform.align2d",t);
01805         }
01806 
01807         delete t;
01808         t = 0;
01809 
01810         gsl_vector_free(x);
01811         gsl_vector_free(ss);
01812         gsl_multimin_fminimizer_free(s);
01813 
01814         if (c != 0) delete c;
01815         return result;
01816 }

virtual string EMAN::RefineAligner::get_desc  )  const [inline, virtual]
 

Implements EMAN::Aligner.

Definition at line 1237 of file aligner.h.

01238                 {
01239                         return "Refines a preliminary 2D alignment using a simplex algorithm. Subpixel precision.";
01240                 }

virtual string EMAN::RefineAligner::get_name  )  const [inline, virtual]
 

Get the Aligner's name.

Each Aligner is identified by a unique name.

Returns:
The Aligner's name.

Implements EMAN::Aligner.

Definition at line 1232 of file aligner.h.

01233                 {
01234                         return NAME;
01235                 }

virtual TypeDict EMAN::RefineAligner::get_param_types  )  const [inline, virtual]
 

Implements EMAN::Aligner.

Definition at line 1247 of file aligner.h.

References EMAN::TypeDict::put().

01248                 {
01249                         TypeDict d;
01250 
01251                         d.put("mode", EMObject::INT, "Currently unused");
01252                         d.put("xform.align2d", EMObject::TRANSFORM, "The Transform storing the starting guess. If unspecified the identity matrix is used");
01253                         d.put("stepx", EMObject::FLOAT, "The x increment used to create the starting simplex. Default is 1");
01254                         d.put("stepy", EMObject::FLOAT, "The y increment used to create the starting simplex. Default is 1");
01255                         d.put("stepaz", EMObject::FLOAT, "The rotational increment used to create the starting simplex. Default is 5");
01256                         d.put("precision", EMObject::FLOAT, "The precision which, if achieved, can stop the iterative refinement before reaching the maximum iterations. Default is 0.04.");
01257                         d.put("maxiter", EMObject::INT,"The maximum number of iterations that can be performed by the Simplex minimizer. default=28");
01258                         d.put("maxshift", EMObject::INT,"Maximum translation in pixels in any direction. If the solution yields a shift beyond this value in any direction, then the refinement is judged a failure and the original alignment is used as the solution.");
01259                         d.put("stepscale", EMObject::FLOAT, "If set to any non-zero value, scale will be included in the alignment, and this will be the initial step. Images should be edgenormalized. If the scale goes beyond +-30% alignment will fail.");
01260                         d.put("mask", EMObject::EMDATA, "A mask to be applied to the image being aligned prior to each similarity comparison.");
01261                         d.put("verbose", EMObject::INT, "This will cause debugging information to be printed on the screen for the iterative refinement. Larger numbers -> more info. default=0");
01262                         return d;
01263                 }

Aligner* EMAN::RefineAligner::NEW  )  [inline, static]
 

Definition at line 1242 of file aligner.h.

01243                 {
01244                         return new RefineAligner();
01245                 }


Member Data Documentation

const string RefineAligner::NAME = "refine" [static]
 

Definition at line 78 of file aligner.cpp.


The documentation for this class was generated from the following files:
Generated on Tue Jun 11 13:47:48 2013 for EMAN2 by  doxygen 1.3.9.1