[ VIGRA Homepage | Function Index | Class Index | Namespaces | File List | Main Page ]

vigra/noise_normalization.hxx VIGRA

00001 /************************************************************************/
00002 /*                                                                      */
00003 /*               Copyright 1998-2006 by Ullrich Koethe                  */
00004 /*                                                                      */
00005 /*    This file is part of the VIGRA computer vision library.           */
00006 /*    The VIGRA Website is                                              */
00007 /*        http://hci.iwr.uni-heidelberg.de/vigra/                       */
00008 /*    Please direct questions, bug reports, and contributions to        */
00009 /*        ullrich.koethe@iwr.uni-heidelberg.de    or                    */
00010 /*        vigra@informatik.uni-hamburg.de                               */
00011 /*                                                                      */
00012 /*    Permission is hereby granted, free of charge, to any person       */
00013 /*    obtaining a copy of this software and associated documentation    */
00014 /*    files (the "Software"), to deal in the Software without           */
00015 /*    restriction, including without limitation the rights to use,      */
00016 /*    copy, modify, merge, publish, distribute, sublicense, and/or      */
00017 /*    sell copies of the Software, and to permit persons to whom the    */
00018 /*    Software is furnished to do so, subject to the following          */
00019 /*    conditions:                                                       */
00020 /*                                                                      */
00021 /*    The above copyright notice and this permission notice shall be    */
00022 /*    included in all copies or substantial portions of the             */
00023 /*    Software.                                                         */
00024 /*                                                                      */
00025 /*    THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND    */
00026 /*    EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES   */
00027 /*    OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND          */
00028 /*    NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT       */
00029 /*    HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY,      */
00030 /*    WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING      */
00031 /*    FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR     */
00032 /*    OTHER DEALINGS IN THE SOFTWARE.                                   */
00033 /*                                                                      */
00034 /************************************************************************/
00035 
00036 
00037 #ifndef VIGRA_NOISE_NORMALIZATION_HXX
00038 #define VIGRA_NOISE_NORMALIZATION_HXX
00039 
00040 #include "utilities.hxx"
00041 #include "tinyvector.hxx"
00042 #include "stdimage.hxx"
00043 #include "transformimage.hxx"
00044 #include "combineimages.hxx"
00045 #include "localminmax.hxx"
00046 #include "functorexpression.hxx"
00047 #include "numerictraits.hxx"
00048 #include "separableconvolution.hxx"
00049 #include "linear_solve.hxx"
00050 #include "array_vector.hxx"
00051 #include "static_assert.hxx"
00052 #include <algorithm>
00053 
00054 namespace vigra {
00055 
00056 /** \addtogroup NoiseNormalization Noise Normalization
00057     Estimate noise with intensity-dependent variance and transform it into additive Gaussian noise.
00058 */
00059 //@{ 
00060                                     
00061 /********************************************************/
00062 /*                                                      */
00063 /*               NoiseNormalizationOptions              */
00064 /*                                                      */
00065 /********************************************************/
00066 
00067 /** \brief Pass options to one of the noise normalization functions.
00068 
00069     <tt>NoiseNormalizationOptions</tt>  is an argument object that holds various optional
00070     parameters used by the noise normalization functions. If a parameter is not explicitly
00071     set, a suitable default will be used.
00072     
00073     <b> Usage:</b>
00074     
00075         <b>\#include</b> <vigra/noise_normalization.hxx><br>
00076     Namespace: vigra
00077     
00078     \code
00079     vigra::BImage src(w,h);
00080     std::vector<vigra::TinyVector<double, 2> > result;
00081     
00082     ...
00083     vigra::noiseVarianceEstimation(srcImageRange(src), result, 
00084                                   vigra::NoiseNormalizationOptions().windowRadius(9).noiseVarianceInitialGuess(25.0));
00085     \endcode
00086 */
00087 class NoiseNormalizationOptions
00088 {
00089   public:
00090   
00091         /** Initialize all options with default values.
00092         */
00093     NoiseNormalizationOptions()
00094     : window_radius(6),
00095       cluster_count(10),
00096       noise_estimation_quantile(1.5),
00097       averaging_quantile(0.8),
00098       noise_variance_initial_guess(10.0),
00099       use_gradient(true)
00100     {}
00101 
00102         /** Select the noise estimation algorithm.
00103         
00104             If \a r is <tt>true</tt>, use the gradient-based noise estimator according to F&ouml;rstner (default).
00105             Otherwise, use an algorithm that uses the intensity values directly.
00106         */
00107     NoiseNormalizationOptions & useGradient(bool r)
00108     {
00109         use_gradient = r;
00110         return *this;
00111     }
00112 
00113         /** Set the window radius for a single noise estimate.
00114             Every window of the given size gives raise to one intensity/variance pair.<br>
00115             Default: 6 pixels
00116         */
00117     NoiseNormalizationOptions & windowRadius(unsigned int r)
00118     {
00119         vigra_precondition(r > 0,
00120             "NoiseNormalizationOptions: window radius must be > 0.");
00121         window_radius = r;
00122         return *this;
00123     }
00124 
00125         /** Set the number of clusters for non-parametric noise normalization.
00126             The intensity/variance pairs found are grouped into clusters before the noise
00127             normalization transform is computed.<br>
00128             Default: 10 clusters
00129         */
00130     NoiseNormalizationOptions & clusterCount(unsigned int c)
00131     {
00132         vigra_precondition(c > 0,
00133             "NoiseNormalizationOptions: cluster count must be > 0.");
00134         cluster_count = c;
00135         return *this;
00136     }
00137 
00138         /** Set the quantile for cluster averaging.
00139             After clustering, the cluster center (i.e. average noise variance as a function of the average
00140             intensity in the cluster) is computed using only the cluster members whose estimated variance
00141             is below \a quantile times the maximum variance in the cluster.<br>
00142             Default: 0.8<br>
00143             Precondition: 0 < \a quantile <= 1.0
00144         */
00145     NoiseNormalizationOptions & averagingQuantile(double quantile)
00146     {
00147         vigra_precondition(quantile > 0.0 && quantile <= 1.0,
00148             "NoiseNormalizationOptions: averaging quantile must be between 0 and 1.");
00149         averaging_quantile = quantile;
00150         return *this;
00151     }
00152 
00153         /** Set the operating range of the robust noise estimator.
00154             Intensity changes that are larger than \a quantile times the current estimate of the noise variance
00155             are ignored by the robust noise estimator.<br>
00156             Default: 1.5<br>
00157             Precondition: 0 < \a quantile
00158         */
00159     NoiseNormalizationOptions & noiseEstimationQuantile(double quantile)
00160     {
00161         vigra_precondition(quantile > 0.0,
00162             "NoiseNormalizationOptions: noise estimation quantile must be > 0.");
00163         noise_estimation_quantile = quantile;
00164         return *this;
00165     }
00166 
00167         /** Set the initial estimate of the noise variance.
00168             Robust noise variance estimation is an iterative procedure starting at the given value.<br>
00169             Default: 10.0<br>
00170             Precondition: 0 < \a guess
00171         */
00172     NoiseNormalizationOptions & noiseVarianceInitialGuess(double guess)
00173     {
00174         vigra_precondition(guess > 0.0,
00175             "NoiseNormalizationOptions: noise variance initial guess must be > 0.");
00176         noise_variance_initial_guess = guess;
00177         return *this;
00178     }
00179 
00180     unsigned int window_radius, cluster_count;
00181     double noise_estimation_quantile, averaging_quantile, noise_variance_initial_guess;
00182     bool use_gradient;
00183 };
00184 
00185 //@}
00186 
00187 template <class ArgumentType, class ResultType>
00188 class NonparametricNoiseNormalizationFunctor
00189 {
00190     struct Segment
00191     {
00192         double lower, a, b, shift;
00193     };
00194 
00195     ArrayVector<Segment> segments_;
00196 
00197     template <class T>
00198     double exec(unsigned int k, T t) const
00199     {
00200         if(segments_[k].a == 0.0)
00201         {
00202             return t / VIGRA_CSTD::sqrt(segments_[k].b);
00203         }
00204         else
00205         {
00206             return 2.0 / segments_[k].a * VIGRA_CSTD::sqrt(std::max(0.0, segments_[k].a * t + segments_[k].b));
00207         }
00208     }
00209 
00210   public:
00211     typedef ArgumentType argument_type;
00212     typedef ResultType result_type;
00213 
00214     template <class Vector>
00215     NonparametricNoiseNormalizationFunctor(Vector const & clusters)
00216     : segments_(clusters.size()-1)
00217     {
00218         for(unsigned int k = 0; k<segments_.size(); ++k)
00219         {
00220             segments_[k].lower = clusters[k][0];
00221             segments_[k].a = (clusters[k+1][1] - clusters[k][1]) / (clusters[k+1][0] - clusters[k][0]);
00222             segments_[k].b = clusters[k][1] - segments_[k].a * clusters[k][0];
00223             // FIXME: set a to zero if it is very small
00224             //          - determine what 'very small' means
00225             //          - shouldn't the two formulas (for a == 0, a != 0) be equal in the limit a -> 0 ?
00226 
00227             if(k == 0)
00228             {
00229                 segments_[k].shift = segments_[k].lower - exec(k, segments_[k].lower);
00230             }
00231             else
00232             {
00233                 segments_[k].shift = exec(k-1, segments_[k].lower) - exec(k, segments_[k].lower) + segments_[k-1].shift;
00234             }
00235         }
00236     }
00237 
00238     result_type operator()(argument_type t) const
00239     {
00240         // find the segment
00241         unsigned int k = 0;
00242         for(; k < segments_.size(); ++k)
00243             if(t < segments_[k].lower)
00244                 break;
00245         if(k > 0)
00246             --k;
00247         return detail::RequiresExplicitCast<ResultType>::cast(exec(k, t) + segments_[k].shift);
00248     }
00249 };
00250 
00251 template <class ArgumentType, class ResultType>
00252 class QuadraticNoiseNormalizationFunctor
00253 {
00254     double a, b, c, d, f, o;
00255 
00256     void init(double ia, double ib, double ic, double xmin)
00257     {
00258         a = ia;
00259         b = ib;
00260         c = ic;
00261         d = VIGRA_CSTD::sqrt(VIGRA_CSTD::fabs(c));
00262         if(c > 0.0)
00263         {
00264             o = VIGRA_CSTD::log(VIGRA_CSTD::fabs((2.0*c*xmin + b)/d + 2*VIGRA_CSTD::sqrt(c*sq(xmin) +b*xmin + a)))/d;
00265             f = 0.0;
00266         }
00267         else
00268         {
00269             f = VIGRA_CSTD::sqrt(b*b - 4.0*a*c);
00270             o = -VIGRA_CSTD::asin((2.0*c*xmin+b)/f)/d;
00271         }
00272     }
00273 
00274   public:
00275     typedef ArgumentType argument_type;
00276     typedef ResultType result_type;
00277 
00278     template <class Vector>
00279     QuadraticNoiseNormalizationFunctor(Vector const & clusters)
00280     {
00281         double xmin = NumericTraits<double>::max();
00282         Matrix<double> m(3,3), r(3, 1), l(3, 1);
00283         for(unsigned int k = 0; k<clusters.size(); ++k)
00284         {
00285             l(0,0) = 1.0;
00286             l(1,0) = clusters[k][0];
00287             l(2,0) = sq(clusters[k][0]);
00288             m += outer(l);
00289             r += clusters[k][1]*l;
00290             if(clusters[k][0] < xmin)
00291                 xmin = clusters[k][0];
00292         }
00293 
00294         linearSolve(m, r, l);
00295         init(l(0,0), l(1,0), l(2,0), xmin);
00296     }
00297 
00298     result_type operator()(argument_type t) const
00299     {
00300         double r;
00301         if(c > 0.0)
00302             r = VIGRA_CSTD::log(VIGRA_CSTD::fabs((2.0*c*t + b)/d + 2.0*VIGRA_CSTD::sqrt(c*t*t +b*t + a)))/d-o;
00303         else
00304             r = -VIGRA_CSTD::asin((2.0*c*t+b)/f)/d-o;
00305         return detail::RequiresExplicitCast<ResultType>::cast(r);
00306     }
00307 };
00308 
00309 template <class ArgumentType, class ResultType>
00310 class LinearNoiseNormalizationFunctor
00311 {
00312     double a, b, o;
00313 
00314     void init(double ia, double ib, double xmin)
00315     {
00316         a = ia;
00317         b = ib;
00318         if(b != 0.0)
00319         {
00320             o = xmin - 2.0 / b * VIGRA_CSTD::sqrt(a + b * xmin);
00321         }
00322         else
00323         {
00324             o = xmin - xmin / VIGRA_CSTD::sqrt(a);
00325         }
00326     }
00327 
00328   public:
00329     typedef ArgumentType argument_type;
00330     typedef ResultType result_type;
00331 
00332     template <class Vector>
00333     LinearNoiseNormalizationFunctor(Vector const & clusters)
00334     {
00335         double xmin = NumericTraits<double>::max();
00336         Matrix<double> m(2,2), r(2, 1), l(2, 1);
00337         for(unsigned int k = 0; k<clusters.size(); ++k)
00338         {
00339             l(0,0) = 1.0;
00340             l(1,0) = clusters[k][0];
00341             m += outer(l);
00342             r += clusters[k][1]*l;
00343             if(clusters[k][0] < xmin)
00344                 xmin = clusters[k][0];
00345         }
00346 
00347         linearSolve(m, r, l);
00348         init(l(0,0), l(1,0), xmin);
00349     }
00350 
00351     result_type operator()(argument_type t) const
00352     {
00353         double r;
00354         if(b != 0.0)
00355             r = 2.0 / b * VIGRA_CSTD::sqrt(a + b*t) + o;
00356         else
00357             r =  t / VIGRA_CSTD::sqrt(a) + o;
00358         return detail::RequiresExplicitCast<ResultType>::cast(r);
00359     }
00360 };
00361 
00362 #define VIGRA_NoiseNormalizationFunctor(name, type, size) \
00363 template <class ResultType> \
00364 class name<type, ResultType> \
00365 { \
00366     ResultType lut_[size]; \
00367     \
00368   public: \
00369     typedef type argument_type; \
00370     typedef ResultType result_type; \
00371      \
00372     template <class Vector> \
00373     name(Vector const & clusters) \
00374     { \
00375         name<double, ResultType> f(clusters); \
00376          \
00377         for(unsigned int k = 0; k < size; ++k) \
00378         { \
00379             lut_[k] = f(k); \
00380         } \
00381     } \
00382      \
00383     result_type operator()(argument_type t) const \
00384     { \
00385         return lut_[t]; \
00386     } \
00387 };
00388 
00389 VIGRA_NoiseNormalizationFunctor(NonparametricNoiseNormalizationFunctor, UInt8, 256)
00390 VIGRA_NoiseNormalizationFunctor(NonparametricNoiseNormalizationFunctor, UInt16, 65536)
00391 VIGRA_NoiseNormalizationFunctor(QuadraticNoiseNormalizationFunctor, UInt8, 256)
00392 VIGRA_NoiseNormalizationFunctor(QuadraticNoiseNormalizationFunctor, UInt16, 65536)
00393 VIGRA_NoiseNormalizationFunctor(LinearNoiseNormalizationFunctor, UInt8, 256)
00394 VIGRA_NoiseNormalizationFunctor(LinearNoiseNormalizationFunctor, UInt16, 65536)
00395 
00396 #undef VIGRA_NoiseNormalizationFunctor
00397 
00398 namespace detail {
00399 
00400 template <class SrcIterator, class SrcAcessor,
00401           class GradIterator>
00402 bool
00403 iterativeNoiseEstimationChi2(SrcIterator s, SrcAcessor src, GradIterator g,
00404                          double & mean, double & variance,
00405                          double robustnessThreshold, int windowRadius)
00406 {
00407     double l2 = sq(robustnessThreshold);
00408     double countThreshold = 1.0 - VIGRA_CSTD::exp(-l2);
00409     double f = (1.0 - VIGRA_CSTD::exp(-l2)) / (1.0 - (1.0 + l2)*VIGRA_CSTD::exp(-l2));
00410 
00411     Diff2D ul(-windowRadius, -windowRadius);
00412     int r2 = sq(windowRadius);
00413 
00414     for(int iter=0; iter<100 ; ++iter) // maximum iteration 100 only for terminating
00415                                        // if something is wrong
00416     {
00417         double sum=0.0;
00418         double gsum=0.0;
00419         unsigned int count = 0;
00420         unsigned int tcount = 0;
00421 
00422         SrcIterator siy = s + ul;
00423         GradIterator giy = g + ul;
00424         for(int y=-windowRadius; y <= windowRadius; y++, ++siy.y, ++giy.y)
00425         {
00426             typename SrcIterator::row_iterator six = siy.rowIterator();
00427             typename GradIterator::row_iterator gix = giy.rowIterator();
00428             for(int x=-windowRadius; x <= windowRadius; x++, ++six, ++gix)
00429             {
00430                 if (sq(x) + sq(y) > r2)
00431                     continue;
00432 
00433                 ++tcount;
00434                 if (*gix < l2*variance)
00435                 {
00436                     sum += src(six);
00437                     gsum += *gix;
00438                     ++count;
00439                 }
00440             }
00441         }
00442         if (count==0) // not homogeneous enough
00443             return false;
00444 
00445         double oldvariance = variance;
00446         variance= f * gsum / count;
00447         mean = sum / count;
00448 
00449         if ( closeAtTolerance(oldvariance - variance, 0.0, 1e-10))
00450             return (count >= tcount * countThreshold / 2.0); // sufficiently many valid points
00451     }
00452     return false; // no convergence
00453 }
00454 
00455 template <class SrcIterator, class SrcAcessor,
00456           class GradIterator>
00457 bool
00458 iterativeNoiseEstimationGauss(SrcIterator s, SrcAcessor src, GradIterator,
00459                          double & mean, double & variance,
00460                          double robustnessThreshold, int windowRadius)
00461 {
00462     double l2 = sq(robustnessThreshold);
00463     double countThreshold = erf(VIGRA_CSTD::sqrt(0.5 * l2));
00464     double f = countThreshold / (countThreshold - VIGRA_CSTD::sqrt(2.0/M_PI*l2)*VIGRA_CSTD::exp(-l2/2.0));
00465 
00466     mean = src(s);
00467 
00468     Diff2D ul(-windowRadius, -windowRadius);
00469     int r2 = sq(windowRadius);
00470 
00471     for(int iter=0; iter<100 ; ++iter) // maximum iteration 100 only for terminating
00472                                        // if something is wrong
00473     {
00474         double sum = 0.0;
00475         double sum2 = 0.0;
00476         unsigned int count = 0;
00477         unsigned int tcount = 0;
00478 
00479         SrcIterator siy = s + ul;
00480         for(int y=-windowRadius; y <= windowRadius; y++, ++siy.y)
00481         {
00482             typename SrcIterator::row_iterator six = siy.rowIterator();
00483             for(int x=-windowRadius; x <= windowRadius; x++, ++six)
00484             {
00485                 if (sq(x) + sq(y) > r2)
00486                     continue;
00487 
00488                 ++tcount;
00489                 if (sq(src(six) - mean) < l2*variance)
00490                 {
00491                     sum += src(six);
00492                     sum2 += sq(src(six));
00493                     ++count;
00494                 }
00495             }
00496         }
00497         if (count==0) // not homogeneous enough
00498             return false;
00499 
00500         double oldmean = mean;
00501         double oldvariance = variance;
00502         mean = sum / count;
00503         variance= f * (sum2 / count - sq(mean));
00504 
00505         if ( closeAtTolerance(oldmean - mean, 0.0, 1e-10) &&
00506              closeAtTolerance(oldvariance - variance, 0.0, 1e-10))
00507             return (count >= tcount * countThreshold / 2.0); // sufficiently many valid points
00508     }
00509     return false; // no convergence
00510 }
00511 
00512 
00513 template <class SrcIterator, class SrcAccessor,
00514           class DestIterator, class DestAccessor>
00515 void
00516 symmetricDifferenceSquaredMagnitude(
00517      SrcIterator sul, SrcIterator slr, SrcAccessor src,
00518      DestIterator dul, DestAccessor dest)
00519 {
00520     using namespace functor;
00521     int w = slr.x - sul.x;
00522     int h = slr.y - sul.y;
00523 
00524     typedef typename NumericTraits<typename SrcAccessor::value_type>::RealPromote TmpType;
00525     typedef BasicImage<TmpType> TmpImage;
00526 
00527     Kernel1D<double> mask;
00528     mask.initSymmetricGradient();
00529     mask.setBorderTreatment(BORDER_TREATMENT_REFLECT);
00530 
00531     TmpImage dx(w, h), dy(w, h);
00532     separableConvolveX(srcIterRange(sul, slr, src), destImage(dx),  kernel1d(mask));
00533     separableConvolveY(srcIterRange(sul, slr, src), destImage(dy),  kernel1d(mask));
00534     combineTwoImages(srcImageRange(dx), srcImage(dy), destIter(dul, dest), Arg1()*Arg1() + Arg2()*Arg2());
00535 }
00536 
00537 template <class SrcIterator, class SrcAccessor,
00538           class DestIterator, class DestAccessor>
00539 void
00540 findHomogeneousRegionsFoerstner(
00541      SrcIterator sul, SrcIterator slr, SrcAccessor src,
00542      DestIterator dul, DestAccessor dest,
00543      unsigned int windowRadius = 6, double homogeneityThreshold = 40.0)
00544 {
00545     using namespace vigra::functor;
00546     int w = slr.x - sul.x;
00547     int h = slr.y - sul.y;
00548 
00549     typedef typename NumericTraits<typename SrcAccessor::value_type>::RealPromote TmpType;
00550     typedef BasicImage<TmpType> TmpImage;
00551 
00552     BImage btmp(w, h);
00553     transformImage(srcIterRange(sul, slr, src), destImage(btmp),
00554                     ifThenElse(Arg1() <= Param(homogeneityThreshold), Param(1), Param(0)));
00555 
00556     // Erosion
00557     discErosion(srcImageRange(btmp), destIter(dul, dest), windowRadius);
00558 }
00559 
00560 template <class SrcIterator, class SrcAccessor,
00561           class DestIterator, class DestAccessor>
00562 void
00563 findHomogeneousRegions(
00564      SrcIterator sul, SrcIterator slr, SrcAccessor src,
00565      DestIterator dul, DestAccessor dest)
00566 {
00567     localMinima(sul, slr, src, dul, dest);
00568 }
00569 
00570 template <class Vector1, class Vector2>
00571 void noiseVarianceListMedianCut(Vector1 const & noise, Vector2 & clusters,
00572                                 unsigned int maxClusterCount)
00573 {
00574     typedef typename Vector2::value_type Result;
00575 
00576     clusters.push_back(Result(0, noise.size()));
00577 
00578     while(clusters.size() <= maxClusterCount)
00579     {
00580         // find biggest cluster
00581         unsigned int kMax = 0;
00582         double diffMax = 0.0;
00583         for(unsigned int k=0; k < clusters.size(); ++k)
00584         {
00585             int k1 = clusters[k][0], k2 = clusters[k][1]-1;
00586             
00587 #if 0       // turned the "internal error" in a postcondition message
00588             // for the most likely case
00589             std::string message("noiseVarianceListMedianCut(): internal error (");
00590             message += std::string("k: ") + asString(k) + ", ";
00591             message += std::string("k1: ") + asString(k1) + ", ";
00592             message += std::string("k2: ") + asString(k2) + ", ";
00593             message += std::string("noise.size(): ") + asString(noise.size()) + ", ";
00594             message += std::string("clusters.size(): ") + asString(clusters.size()) + ").";
00595             vigra_invariant(k1 >= 0 && k1 < (int)noise.size() && k2 >= 0 && k2 < (int)noise.size(), message.c_str());
00596 #endif
00597             
00598             vigra_postcondition(k1 >= 0 && k1 < (int)noise.size() && 
00599                                 k2 >= 0 && k2 < (int)noise.size(), 
00600                 "noiseVarianceClustering(): Unable to find homogeneous regions.");
00601 
00602             double diff = noise[k2][0] - noise[k1][0];
00603             if(diff > diffMax)
00604             {
00605                 diffMax = diff;
00606                 kMax = k;
00607             }
00608         }
00609 
00610         if(diffMax == 0.0)
00611             return; // all clusters have only one value
00612 
00613         unsigned int k1 = clusters[kMax][0],
00614                      k2 = clusters[kMax][1];
00615         unsigned int kSplit = k1 + (k2 - k1) / 2;
00616         clusters[kMax][1] = kSplit;
00617         clusters.push_back(Result(kSplit, k2));
00618     }
00619 }
00620 
00621 struct SortNoiseByMean
00622 {
00623     template <class T>
00624     bool operator()(T const & l, T const & r) const
00625     {
00626         return l[0] < r[0];
00627     }
00628 };
00629 
00630 struct SortNoiseByVariance
00631 {
00632     template <class T>
00633     bool operator()(T const & l, T const & r) const
00634     {
00635         return l[1] < r[1];
00636     }
00637 };
00638 
00639 template <class Vector1, class Vector2, class Vector3>
00640 void noiseVarianceClusterAveraging(Vector1 & noise, Vector2 & clusters,
00641                                    Vector3 & result, double quantile)
00642 {
00643     typedef typename Vector1::iterator Iter;
00644     typedef typename Vector3::value_type Result;
00645 
00646     for(unsigned int k=0; k<clusters.size(); ++k)
00647     {
00648         Iter i1 = noise.begin() + clusters[k][0];
00649         Iter i2 = noise.begin() + clusters[k][1];
00650 
00651         std::sort(i1, i2, SortNoiseByVariance());
00652 
00653         std::size_t size = static_cast<std::size_t>(VIGRA_CSTD::ceil(quantile*(i2 - i1)));
00654         if(static_cast<std::size_t>(i2 - i1) < size)
00655             size = i2 - i1;
00656         if(size < 1)
00657             size = 1;
00658         i2 = i1 + size;
00659 
00660         double mean = 0.0,
00661                variance = 0.0;
00662         for(; i1 < i2; ++i1)
00663         {
00664             mean += (*i1)[0];
00665             variance += (*i1)[1];
00666         }
00667 
00668         result.push_back(Result(mean / size, variance / size));
00669     }
00670 }
00671 
00672 template <class SrcIterator, class SrcAccessor, class BackInsertable>
00673 void noiseVarianceEstimationImpl(SrcIterator sul, SrcIterator slr, SrcAccessor src,
00674                            BackInsertable & result,
00675                            NoiseNormalizationOptions const & options)
00676 {
00677     typedef typename BackInsertable::value_type ResultType;
00678 
00679     unsigned int w = slr.x - sul.x;
00680     unsigned int h = slr.y - sul.y;
00681 
00682     typedef typename NumericTraits<typename SrcAccessor::value_type>::RealPromote TmpType;
00683     typedef BasicImage<TmpType> TmpImage;
00684 
00685     TmpImage gradient(w, h);
00686     symmetricDifferenceSquaredMagnitude(sul, slr, src, gradient.upperLeft(), gradient.accessor());
00687 
00688     BImage homogeneous(w, h);
00689     findHomogeneousRegions(gradient.upperLeft(), gradient.lowerRight(), gradient.accessor(),
00690                                    homogeneous.upperLeft(), homogeneous.accessor());
00691 
00692     // Generate noise of each of the remaining pixels == centers of homogeneous areas (border is not used)
00693     unsigned int windowRadius = options.window_radius;
00694     for(unsigned int y=windowRadius; y<h-windowRadius; ++y)
00695     {
00696         for(unsigned int x=windowRadius; x<w-windowRadius; ++x)
00697         {
00698             if (! homogeneous(x, y))
00699                 continue;
00700 
00701             Diff2D center(x, y);
00702             double mean = 0.0, variance = options.noise_variance_initial_guess;
00703 
00704             bool success;
00705 
00706             if(options.use_gradient)
00707             {
00708                 success = iterativeNoiseEstimationChi2(sul + center, src,
00709                               gradient.upperLeft() + center, mean, variance,
00710                               options.noise_estimation_quantile, windowRadius);
00711             }
00712             else
00713             {
00714                 success = iterativeNoiseEstimationGauss(sul + center, src,
00715                               gradient.upperLeft() + center, mean, variance,
00716                               options.noise_estimation_quantile, windowRadius);
00717             }
00718             if (success)
00719             {
00720                 result.push_back(ResultType(mean, variance));
00721             }
00722         }
00723     }
00724 }
00725 
00726 template <class Vector, class BackInsertable>
00727 void noiseVarianceClusteringImpl(Vector & noise, BackInsertable & result,
00728                            unsigned int clusterCount, double quantile)
00729 {
00730     std::sort(noise.begin(), noise.end(), detail::SortNoiseByMean());
00731 
00732     ArrayVector<TinyVector<unsigned int, 2> > clusters;
00733     detail::noiseVarianceListMedianCut(noise, clusters, clusterCount);
00734 
00735     std::sort(clusters.begin(), clusters.end(), detail::SortNoiseByMean());
00736 
00737     detail::noiseVarianceClusterAveraging(noise, clusters, result, quantile);
00738 }
00739 
00740 template <class Functor,
00741           class SrcIterator, class SrcAccessor,
00742           class DestIterator, class DestAccessor>
00743 bool
00744 noiseNormalizationImpl(SrcIterator sul, SrcIterator slr, SrcAccessor src,
00745                        DestIterator dul, DestAccessor dest,
00746                        NoiseNormalizationOptions const & options)
00747 {
00748     ArrayVector<TinyVector<double, 2> > noiseData;
00749     noiseVarianceEstimationImpl(sul, slr, src, noiseData, options);
00750 
00751     if(noiseData.size() < 10)
00752         return false;
00753 
00754     ArrayVector<TinyVector<double, 2> > noiseClusters;
00755 
00756     noiseVarianceClusteringImpl(noiseData, noiseClusters,
00757                                   options.cluster_count, options.averaging_quantile);
00758 
00759     transformImage(sul, slr, src, dul, dest, Functor(noiseClusters));
00760 
00761     return true;
00762 }
00763 
00764 template <class SrcIterator, class SrcAccessor,
00765           class DestIterator, class DestAccessor>
00766 bool
00767 nonparametricNoiseNormalizationImpl(SrcIterator sul, SrcIterator slr, SrcAccessor src,
00768                                     DestIterator dul, DestAccessor dest,
00769                                     NoiseNormalizationOptions const & options,
00770                                     VigraTrueType /* isScalar */)
00771 {
00772     typedef typename SrcAccessor::value_type SrcType;
00773     typedef typename DestAccessor::value_type DestType;
00774     return noiseNormalizationImpl<NonparametricNoiseNormalizationFunctor<SrcType, DestType> >
00775                                                          (sul, slr, src, dul, dest, options);
00776 }
00777 
00778 template <class SrcIterator, class SrcAccessor,
00779           class DestIterator, class DestAccessor>
00780 bool
00781 nonparametricNoiseNormalizationImpl(SrcIterator sul, SrcIterator slr, SrcAccessor src,
00782                               DestIterator dul, DestAccessor dest,
00783                               NoiseNormalizationOptions const & options,
00784                               VigraFalseType /* isScalar */)
00785 {
00786     int bands = src.size(sul);
00787     for(int b=0; b<bands; ++b)
00788     {
00789         VectorElementAccessor<SrcAccessor> sband(b, src);
00790         VectorElementAccessor<DestAccessor> dband(b, dest);
00791         typedef typename VectorElementAccessor<SrcAccessor>::value_type SrcType;
00792         typedef typename VectorElementAccessor<DestAccessor>::value_type DestType;
00793 
00794         if(!noiseNormalizationImpl<NonparametricNoiseNormalizationFunctor<SrcType, DestType> >
00795                                                            (sul, slr, sband, dul, dband, options))
00796             return false;
00797     }
00798     return true;
00799 }
00800 
00801 template <class SrcIterator, class SrcAccessor,
00802           class DestIterator, class DestAccessor>
00803 bool
00804 quadraticNoiseNormalizationImpl(SrcIterator sul, SrcIterator slr, SrcAccessor src,
00805                                     DestIterator dul, DestAccessor dest,
00806                                     NoiseNormalizationOptions const & options,
00807                                     VigraTrueType /* isScalar */)
00808 {
00809     typedef typename SrcAccessor::value_type SrcType;
00810     typedef typename DestAccessor::value_type DestType;
00811     return noiseNormalizationImpl<QuadraticNoiseNormalizationFunctor<SrcType, DestType> >
00812                                                          (sul, slr, src, dul, dest, options);
00813 }
00814 
00815 template <class SrcIterator, class SrcAccessor,
00816           class DestIterator, class DestAccessor>
00817 bool
00818 quadraticNoiseNormalizationImpl(SrcIterator sul, SrcIterator slr, SrcAccessor src,
00819                               DestIterator dul, DestAccessor dest,
00820                               NoiseNormalizationOptions const & options,
00821                               VigraFalseType /* isScalar */)
00822 {
00823     int bands = src.size(sul);
00824     for(int b=0; b<bands; ++b)
00825     {
00826         VectorElementAccessor<SrcAccessor> sband(b, src);
00827         VectorElementAccessor<DestAccessor> dband(b, dest);
00828         typedef typename VectorElementAccessor<SrcAccessor>::value_type SrcType;
00829         typedef typename VectorElementAccessor<DestAccessor>::value_type DestType;
00830 
00831         if(!noiseNormalizationImpl<QuadraticNoiseNormalizationFunctor<SrcType, DestType> >
00832                                                            (sul, slr, sband, dul, dband, options))
00833             return false;
00834     }
00835     return true;
00836 }
00837 
00838 template <class SrcIterator, class SrcAccessor,
00839           class DestIterator, class DestAccessor>
00840 void
00841 quadraticNoiseNormalizationImpl(SrcIterator sul, SrcIterator slr, SrcAccessor src,
00842                               DestIterator dul, DestAccessor dest,
00843                               double a0, double a1, double a2,
00844                               VigraTrueType /* isScalar */)
00845 {
00846     ArrayVector<TinyVector<double, 2> > noiseClusters;
00847     noiseClusters.push_back(TinyVector<double, 2>(0.0, a0));
00848     noiseClusters.push_back(TinyVector<double, 2>(1.0, a0 + a1 + a2));
00849     noiseClusters.push_back(TinyVector<double, 2>(2.0, a0 + 2.0*a1 + 4.0*a2));
00850     transformImage(sul, slr, src, dul, dest,
00851                    QuadraticNoiseNormalizationFunctor<typename SrcAccessor::value_type,
00852                                                    typename DestAccessor::value_type>(noiseClusters));
00853 }
00854 
00855 template <class SrcIterator, class SrcAccessor,
00856           class DestIterator, class DestAccessor>
00857 void
00858 quadraticNoiseNormalizationImpl(SrcIterator sul, SrcIterator slr, SrcAccessor src,
00859                               DestIterator dul, DestAccessor dest,
00860                               double a0, double a1, double a2,
00861                               VigraFalseType /* isScalar */)
00862 {
00863     int bands = src.size(sul);
00864     for(int b=0; b<bands; ++b)
00865     {
00866         VectorElementAccessor<SrcAccessor> sband(b, src);
00867         VectorElementAccessor<DestAccessor> dband(b, dest);
00868         quadraticNoiseNormalizationImpl(sul, slr, sband, dul, dband, a0, a1, a2, VigraTrueType());
00869     }
00870 }
00871 
00872 template <class SrcIterator, class SrcAccessor,
00873           class DestIterator, class DestAccessor>
00874 bool
00875 linearNoiseNormalizationImpl(SrcIterator sul, SrcIterator slr, SrcAccessor src,
00876                                     DestIterator dul, DestAccessor dest,
00877                                     NoiseNormalizationOptions const & options,
00878                                     VigraTrueType /* isScalar */)
00879 {
00880     typedef typename SrcAccessor::value_type SrcType;
00881     typedef typename DestAccessor::value_type DestType;
00882     return noiseNormalizationImpl<LinearNoiseNormalizationFunctor<SrcType, DestType> >
00883                                                          (sul, slr, src, dul, dest, options);
00884 }
00885 
00886 template <class SrcIterator, class SrcAccessor,
00887           class DestIterator, class DestAccessor>
00888 bool
00889 linearNoiseNormalizationImpl(SrcIterator sul, SrcIterator slr, SrcAccessor src,
00890                               DestIterator dul, DestAccessor dest,
00891                               NoiseNormalizationOptions const & options,
00892                               VigraFalseType /* isScalar */)
00893 {
00894     int bands = src.size(sul);
00895     for(int b=0; b<bands; ++b)
00896     {
00897         VectorElementAccessor<SrcAccessor> sband(b, src);
00898         VectorElementAccessor<DestAccessor> dband(b, dest);
00899         typedef typename VectorElementAccessor<SrcAccessor>::value_type SrcType;
00900         typedef typename VectorElementAccessor<DestAccessor>::value_type DestType;
00901 
00902         if(!noiseNormalizationImpl<LinearNoiseNormalizationFunctor<SrcType, DestType> >
00903                                                            (sul, slr, sband, dul, dband, options))
00904             return false;
00905     }
00906     return true;
00907 }
00908 
00909 template <class SrcIterator, class SrcAccessor,
00910           class DestIterator, class DestAccessor>
00911 void
00912 linearNoiseNormalizationImpl(SrcIterator sul, SrcIterator slr, SrcAccessor src,
00913                               DestIterator dul, DestAccessor dest,
00914                               double a0, double a1,
00915                               VigraTrueType /* isScalar */)
00916 {
00917     ArrayVector<TinyVector<double, 2> > noiseClusters;
00918     noiseClusters.push_back(TinyVector<double, 2>(0.0, a0));
00919     noiseClusters.push_back(TinyVector<double, 2>(1.0, a0 + a1));
00920     transformImage(sul, slr, src, dul, dest,
00921                    LinearNoiseNormalizationFunctor<typename SrcAccessor::value_type,
00922                                                    typename DestAccessor::value_type>(noiseClusters));
00923 }
00924 
00925 template <class SrcIterator, class SrcAccessor,
00926           class DestIterator, class DestAccessor>
00927 void
00928 linearNoiseNormalizationImpl(SrcIterator sul, SrcIterator slr, SrcAccessor src,
00929                               DestIterator dul, DestAccessor dest,
00930                               double a0, double a1,
00931                               VigraFalseType /* isScalar */)
00932 {
00933     int bands = src.size(sul);
00934     for(int b=0; b<bands; ++b)
00935     {
00936         VectorElementAccessor<SrcAccessor> sband(b, src);
00937         VectorElementAccessor<DestAccessor> dband(b, dest);
00938         linearNoiseNormalizationImpl(sul, slr, sband, dul, dband, a0, a1, VigraTrueType());
00939     }
00940 }
00941 
00942 } // namespace detail
00943 
00944 template <bool P>
00945 struct noiseVarianceEstimation_can_only_work_on_scalar_images
00946 : vigra::staticAssert::AssertBool<P>
00947 {};
00948 
00949 /** \addtogroup NoiseNormalization Noise Normalization
00950     Estimate noise with intensity-dependent variance and transform it into additive Gaussian noise.
00951 */
00952 //@{ 
00953                                     
00954 /********************************************************/
00955 /*                                                      */
00956 /*                noiseVarianceEstimation               */
00957 /*                                                      */
00958 /********************************************************/
00959 
00960 /** \brief Determine the noise variance as a function of the image intensity.
00961 
00962     This operator applies an algorithm described in 
00963     
00964     W. F&ouml;rstner: <i>"Image Preprocessing for Feature Extraction in Digital Intensity, Color and Range Images"</i>, 
00965     Proc. Summer School on Data Analysis and the Statistical Foundations of Geomatics, 
00966     Lecture Notes in Earth Science, Berlin: Springer, 1999
00967     
00968     in order to estimate the noise variance as a function of the image intensity in a robust way,
00969     i.e. so that intensity changes due to edges do not bias the estimate. The source value type 
00970     (<TT>SrcAccessor::value_type</TT>) must be a scalar type which is convertible to <tt>double</tt>.
00971     The result is written into the \a result sequence, whose <tt>value_type</tt> must be constructible 
00972     from two <tt>double</tt> values. The following options can be set via the \a options object 
00973     (see \ref vigra::NoiseNormalizationOptions for details):<br><br>
00974     
00975     <tt>useGradient</tt>, <tt>windowRadius</tt>, <tt>noiseEstimationQuantile</tt>, <tt>noiseVarianceInitialGuess</tt>
00976     
00977     <b> Declarations:</b>
00978     
00979     pass arguments explicitly:
00980     \code
00981     namespace vigra {
00982         template <class SrcIterator, class SrcAccessor, class BackInsertable>
00983         void noiseVarianceEstimation(SrcIterator sul, SrcIterator slr, SrcAccessor src,
00984                                      BackInsertable & result,
00985                                      NoiseNormalizationOptions const & options = NoiseNormalizationOptions());
00986     }
00987     \endcode
00988     
00989     use argument objects in conjunction with \ref ArgumentObjectFactories :
00990     \code
00991     namespace vigra {
00992         template <class SrcIterator, class SrcAccessor, class BackInsertable>
00993         void noiseVarianceEstimation(triple<SrcIterator, SrcIterator, SrcAccessor> src,
00994                                      BackInsertable & result,
00995                                      NoiseNormalizationOptions const & options = NoiseNormalizationOptions());
00996     }
00997     \endcode
00998     
00999     <b> Usage:</b>
01000     
01001         <b>\#include</b> <vigra/noise_normalization.hxx><br>
01002     Namespace: vigra
01003     
01004     \code
01005     vigra::BImage src(w,h);
01006     std::vector<vigra::TinyVector<double, 2> > result;
01007     
01008     ...
01009     vigra::noiseVarianceEstimation(srcImageRange(src), result, 
01010                                   vigra::NoiseNormalizationOptions().windowRadius(9).noiseVarianceInitialGuess(25.0));
01011     
01012     // print the intensity / variance pairs found
01013     for(int k=0; k<result.size(); ++k)
01014         std::cout << "Intensity: " << result[k][0] << ", estimated variance: " << result[k][1] << std::endl;
01015     \endcode
01016 
01017     <b> Required Interface:</b>
01018     
01019     \code
01020     SrcIterator upperleft, lowerright;
01021     SrcAccessor src;
01022     
01023     typedef SrcAccessor::value_type SrcType;
01024     typedef NumericTraits<SrcType>::isScalar isScalar;
01025     assert(isScalar::asBool == true);
01026     
01027     double value = src(uperleft);
01028     
01029     BackInsertable result;
01030     typedef BackInsertable::value_type ResultType;    
01031     double intensity, variance;
01032     result.push_back(ResultType(intensity, variance));
01033     \endcode
01034 */
01035 doxygen_overloaded_function(template <...> void noiseVarianceEstimation)
01036 
01037 template <class SrcIterator, class SrcAccessor, class BackInsertable>
01038 inline
01039 void noiseVarianceEstimation(SrcIterator sul, SrcIterator slr, SrcAccessor src,
01040                            BackInsertable & result,
01041                            NoiseNormalizationOptions const & options = NoiseNormalizationOptions())
01042 {
01043     typedef typename BackInsertable::value_type ResultType;
01044     typedef typename SrcAccessor::value_type SrcType;
01045     typedef typename NumericTraits<SrcType>::isScalar isScalar;
01046 
01047     VIGRA_STATIC_ASSERT((
01048         noiseVarianceEstimation_can_only_work_on_scalar_images<(isScalar::asBool)>));
01049 
01050     detail::noiseVarianceEstimationImpl(sul, slr, src, result, options);
01051 }
01052 
01053 template <class SrcIterator, class SrcAccessor, class BackInsertable>
01054 inline
01055 void noiseVarianceEstimation(triple<SrcIterator, SrcIterator, SrcAccessor> src,
01056                            BackInsertable & result,
01057                            NoiseNormalizationOptions const & options = NoiseNormalizationOptions())
01058 {
01059     noiseVarianceEstimation(src.first, src.second, src.third, result, options);
01060 }
01061 
01062 /********************************************************/
01063 /*                                                      */
01064 /*                noiseVarianceClustering               */
01065 /*                                                      */
01066 /********************************************************/
01067 
01068 /** \brief Determine the noise variance as a function of the image intensity and cluster the results.
01069 
01070     This operator first calls \ref noiseVarianceEstimation() to obtain a sequence of intensity/variance pairs,
01071     which are then clustered using the median cut algorithm. Then the cluster centers (i.e. average variance vs.
01072     average intensity) are determined and returned in the \a result sequence.
01073     
01074     In addition to the options valid for \ref noiseVarianceEstimation(), the following options can be set via 
01075     the \a options object (see \ref vigra::NoiseNormalizationOptions for details):<br><br>
01076     
01077     <tt>clusterCount</tt>, <tt>averagingQuantile</tt>
01078     
01079     <b> Declarations:</b>
01080     
01081     pass arguments explicitly:
01082     \code
01083     namespace vigra {
01084         template <class SrcIterator, class SrcAccessor, class BackInsertable>
01085         void noiseVarianceClustering(SrcIterator sul, SrcIterator slr, SrcAccessor src,
01086                                 BackInsertable & result,
01087                                 NoiseNormalizationOptions const & options = NoiseNormalizationOptions());
01088     }
01089     \endcode
01090     
01091     use argument objects in conjunction with \ref ArgumentObjectFactories :
01092     \code
01093     namespace vigra {
01094         template <class SrcIterator, class SrcAccessor, class BackInsertable>
01095         void noiseVarianceClustering(triple<SrcIterator, SrcIterator, SrcAccessor> src,
01096                                 BackInsertable & result,
01097                                 NoiseNormalizationOptions const & options = NoiseNormalizationOptions());
01098     }
01099     \endcode
01100     
01101     <b> Usage:</b>
01102     
01103         <b>\#include</b> <vigra/noise_normalization.hxx><br>
01104     Namespace: vigra
01105     
01106     \code
01107     vigra::BImage src(w,h);
01108     std::vector<vigra::TinyVector<double, 2> > result;
01109     
01110     ...
01111     vigra::noiseVarianceClustering(srcImageRange(src), result, 
01112                                   vigra::NoiseNormalizationOptions().windowRadius(9).noiseVarianceInitialGuess(25.0).
01113                                   clusterCount(15));
01114     
01115     // print the intensity / variance pairs representing the cluster centers
01116     for(int k=0; k<result.size(); ++k)
01117         std::cout << "Cluster: " << k << ", intensity: " << result[k][0] << ", estimated variance: " << result[k][1] << std::endl;
01118     \endcode
01119 
01120     <b> Required Interface:</b>
01121     
01122     same as \ref noiseVarianceEstimation()
01123 */
01124 doxygen_overloaded_function(template <...> void noiseVarianceClustering)
01125 
01126 template <class SrcIterator, class SrcAccessor, class BackInsertable>
01127 inline
01128 void noiseVarianceClustering(SrcIterator sul, SrcIterator slr, SrcAccessor src,
01129                            BackInsertable & result,
01130                            NoiseNormalizationOptions const & options = NoiseNormalizationOptions())
01131 {
01132     ArrayVector<TinyVector<double, 2> > variance;
01133     noiseVarianceEstimation(sul, slr, src, variance, options);
01134     detail::noiseVarianceClusteringImpl(variance, result, options.cluster_count, options.averaging_quantile);
01135 }
01136 
01137 template <class SrcIterator, class SrcAccessor, class BackInsertable>
01138 inline
01139 void noiseVarianceClustering(triple<SrcIterator, SrcIterator, SrcAccessor> src,
01140                            BackInsertable & result,
01141                            NoiseNormalizationOptions const & options = NoiseNormalizationOptions())
01142 {
01143     noiseVarianceClustering(src.first, src.second, src.third, result, options);
01144 }
01145 
01146 /********************************************************/
01147 /*                                                      */
01148 /*             nonparametricNoiseNormalization          */
01149 /*                                                      */
01150 /********************************************************/
01151 
01152 /** \brief Noise normalization by means of an estimated non-parametric noise model.
01153 
01154     The original image is assumed to be corrupted by noise whose variance depends on the intensity in an unknown way.
01155     The present functions first calls \ref noiseVarianceClustering() to obtain a sequence of intensity/variance pairs
01156     (cluster centers) which estimate this dependency. The cluster centers are connected into a piecewise linear 
01157     function which is the inverted according to the formula derived in 
01158     
01159     W. F&ouml;rstner: <i>"Image Preprocessing for Feature Extraction in Digital Intensity, Color and Range Images"</i>, 
01160     Proc. Summer School on Data Analysis and the Statistical Foundations of Geomatics, 
01161     Lecture Notes in Earth Science, Berlin: Springer, 1999
01162 
01163     The inverted formula defines a pixel-wise intensity transformation whose application turns the original image
01164     into one that is corrupted by additive Gaussian noise with unit variance. Most subsequent algorithms will be able
01165     to handle this type of noise much better than the original noise.
01166     
01167     RGB and other multiband images will be processed one band at a time. The function returns <tt>true</tt> on success.
01168     Noise normalization will fail if the original image does not contain sufficiently homogeneous regions that
01169     allow robust estimation of the noise variance.
01170     
01171     The \a options object may use all options described in \ref vigra::NoiseNormalizationOptions.
01172     
01173     <b> Declarations:</b>
01174     
01175     pass arguments explicitly:
01176     \code
01177     namespace vigra {
01178         template <class SrcIterator, class SrcAccessor,
01179                   class DestIterator, class DestAccessor>
01180         bool nonparametricNoiseNormalization(SrcIterator sul, SrcIterator slr, SrcAccessor src,
01181                                              DestIterator dul, DestAccessor dest,
01182                                              NoiseNormalizationOptions const & options = NoiseNormalizationOptions());
01183     }
01184     \endcode
01185     
01186     use argument objects in conjunction with \ref ArgumentObjectFactories :
01187     \code
01188     namespace vigra {
01189         template <class SrcIterator, class SrcAccessor,
01190                   class DestIterator, class DestAccessor>
01191         bool nonparametricNoiseNormalization(triple<SrcIterator, SrcIterator, SrcAccessor> src,
01192                                              pair<DestIterator, DestAccessor> dest,
01193                                              NoiseNormalizationOptions const & options = NoiseNormalizationOptions());
01194     }
01195     \endcode
01196     
01197     <b> Usage:</b>
01198     
01199         <b>\#include</b> <vigra/noise_normalization.hxx><br>
01200     Namespace: vigra
01201     
01202     \code
01203     vigra::BRGBImage src(w,h), dest(w, h);
01204     
01205     ...
01206     vigra::nonparametricNoiseNormalization(srcImageRange(src), destImage(dest), 
01207                                            vigra::NoiseNormalizationOptions().windowRadius(9).noiseVarianceInitialGuess(25.0).
01208                                            clusterCount(15));
01209     \endcode
01210 
01211     <b> Required Interface:</b>
01212     
01213     same as \ref noiseVarianceEstimation()
01214 */
01215 doxygen_overloaded_function(template <...> bool nonparametricNoiseNormalization)
01216 
01217 template <class SrcIterator, class SrcAccessor,
01218           class DestIterator, class DestAccessor>
01219 inline bool
01220 nonparametricNoiseNormalization(SrcIterator sul, SrcIterator slr, SrcAccessor src,
01221                                 DestIterator dul, DestAccessor dest,
01222                                 NoiseNormalizationOptions const & options = NoiseNormalizationOptions())
01223 {
01224     typedef typename SrcAccessor::value_type SrcType;
01225 
01226     return detail::nonparametricNoiseNormalizationImpl(sul, slr, src, dul, dest, options,
01227                                          typename NumericTraits<SrcType>::isScalar());
01228 }
01229 
01230 template <class SrcIterator, class SrcAccessor,
01231           class DestIterator, class DestAccessor>
01232 inline
01233 bool nonparametricNoiseNormalization(triple<SrcIterator, SrcIterator, SrcAccessor> src,
01234                                      pair<DestIterator, DestAccessor> dest,
01235                                      NoiseNormalizationOptions const & options = NoiseNormalizationOptions())
01236 {
01237     return nonparametricNoiseNormalization(src.first, src.second, src.third, dest.first, dest.second, options);
01238 }
01239 
01240 /********************************************************/
01241 /*                                                      */
01242 /*               quadraticNoiseNormalization            */
01243 /*                                                      */
01244 /********************************************************/
01245 
01246 /** \brief Noise normalization by means of an estimated quadratic noise model.
01247 
01248     This function works in the same way as \ref nonparametricNoiseNormalization() with the exception of the 
01249     model for the dependency between intensity and noise variance: it assumes that this dependency is a 
01250     quadratic function rather than a piecewise linear function. If the quadratic model is applicable, it leads
01251     to a somewhat smoother transformation.
01252     
01253     <b> Declarations:</b>
01254     
01255     pass arguments explicitly:
01256     \code
01257     namespace vigra {
01258         template <class SrcIterator, class SrcAccessor,
01259                   class DestIterator, class DestAccessor>
01260         bool quadraticNoiseNormalization(SrcIterator sul, SrcIterator slr, SrcAccessor src,
01261                                          DestIterator dul, DestAccessor dest,
01262                                          NoiseNormalizationOptions const & options = NoiseNormalizationOptions());
01263     }
01264     \endcode
01265     
01266     use argument objects in conjunction with \ref ArgumentObjectFactories :
01267     \code
01268     namespace vigra {
01269         template <class SrcIterator, class SrcAccessor,
01270                   class DestIterator, class DestAccessor>
01271         bool quadraticNoiseNormalization(triple<SrcIterator, SrcIterator, SrcAccessor> src,
01272                                          pair<DestIterator, DestAccessor> dest,
01273                                          NoiseNormalizationOptions const & options = NoiseNormalizationOptions());
01274     }
01275     \endcode
01276     
01277     <b> Usage:</b>
01278     
01279         <b>\#include</b> <vigra/noise_normalization.hxx><br>
01280     Namespace: vigra
01281     
01282     \code
01283     vigra::BRGBImage src(w,h), dest(w, h);
01284     
01285     ...
01286     vigra::quadraticNoiseNormalization(srcImageRange(src), destImage(dest), 
01287                                        vigra::NoiseNormalizationOptions().windowRadius(9).noiseVarianceInitialGuess(25.0).
01288                                        clusterCount(15));
01289     \endcode
01290 
01291     <b> Required Interface:</b>
01292     
01293     same as \ref noiseVarianceEstimation()
01294 */
01295 doxygen_overloaded_function(template <...> bool quadraticNoiseNormalization)
01296 
01297 template <class SrcIterator, class SrcAccessor,
01298           class DestIterator, class DestAccessor>
01299 inline bool
01300 quadraticNoiseNormalization(SrcIterator sul, SrcIterator slr, SrcAccessor src,
01301                                 DestIterator dul, DestAccessor dest,
01302                                 NoiseNormalizationOptions const & options = NoiseNormalizationOptions())
01303 {
01304     typedef typename SrcAccessor::value_type SrcType;
01305 
01306     return detail::quadraticNoiseNormalizationImpl(sul, slr, src, dul, dest, options,
01307                                          typename NumericTraits<SrcType>::isScalar());
01308 }
01309 
01310 template <class SrcIterator, class SrcAccessor,
01311           class DestIterator, class DestAccessor>
01312 inline
01313 bool quadraticNoiseNormalization(triple<SrcIterator, SrcIterator, SrcAccessor> src,
01314                                      pair<DestIterator, DestAccessor> dest,
01315                                      NoiseNormalizationOptions const & options = NoiseNormalizationOptions())
01316 {
01317     return quadraticNoiseNormalization(src.first, src.second, src.third, dest.first, dest.second, options);
01318 }
01319 
01320 /********************************************************/
01321 /*                                                      */
01322 /*               quadraticNoiseNormalization            */
01323 /*                                                      */
01324 /********************************************************/
01325 
01326 /** \brief Noise normalization by means of a given quadratic noise model.
01327 
01328     This function works similar to \ref nonparametricNoiseNormalization() with the exception that the 
01329     functional dependency of the noise variance from the intensity is given (by a quadratic function)
01330     rather than estimated:
01331     
01332     \code
01333     variance = a0 + a1 * intensity + a2 * sq(intensity)
01334     \endcode
01335     
01336     <b> Declarations:</b>
01337     
01338     pass arguments explicitly:
01339     \code
01340     namespace vigra {
01341         template <class SrcIterator, class SrcAccessor,
01342                   class DestIterator, class DestAccessor>
01343         void quadraticNoiseNormalization(SrcIterator sul, SrcIterator slr, SrcAccessor src,
01344                                          DestIterator dul, DestAccessor dest,
01345                                          double a0, double a1, double a2);
01346     }
01347     \endcode
01348     
01349     use argument objects in conjunction with \ref ArgumentObjectFactories :
01350     \code
01351     namespace vigra {
01352         template <class SrcIterator, class SrcAccessor,
01353                   class DestIterator, class DestAccessor>
01354         void quadraticNoiseNormalization(triple<SrcIterator, SrcIterator, SrcAccessor> src,
01355                                         pair<DestIterator, DestAccessor> dest,
01356                                         double a0, double a1, double a2);
01357     }
01358     \endcode
01359     
01360     <b> Usage:</b>
01361     
01362         <b>\#include</b> <vigra/noise_normalization.hxx><br>
01363     Namespace: vigra
01364     
01365     \code
01366     vigra::BRGBImage src(w,h), dest(w, h);
01367     
01368     ...
01369     vigra::quadraticNoiseNormalization(srcImageRange(src), destImage(dest), 
01370                                        100, 0.02, 1e-6);
01371     \endcode
01372 
01373     <b> Required Interface:</b>
01374     
01375     The source value type must be convertible to <tt>double</tt> or must be a vector whose elements 
01376     are convertible to <tt>double</tt>. Likewise, the destination type must be assignable from <tt>double</tt>
01377     or a vector whose elements are assignable from <tt>double</tt>.
01378 */
01379 doxygen_overloaded_function(template <...> void quadraticNoiseNormalization)
01380 
01381 template <class SrcIterator, class SrcAccessor,
01382           class DestIterator, class DestAccessor>
01383 inline void
01384 quadraticNoiseNormalization(SrcIterator sul, SrcIterator slr, SrcAccessor src,
01385                             DestIterator dul, DestAccessor dest,
01386                             double a0, double a1, double a2)
01387 {
01388     typedef typename SrcAccessor::value_type SrcType;
01389 
01390     detail::quadraticNoiseNormalizationImpl(sul, slr, src, dul, dest, a0, a1, a2,
01391                                          typename NumericTraits<SrcType>::isScalar());
01392 }
01393 
01394 template <class SrcIterator, class SrcAccessor,
01395           class DestIterator, class DestAccessor>
01396 inline
01397 void quadraticNoiseNormalization(triple<SrcIterator, SrcIterator, SrcAccessor> src,
01398                                  pair<DestIterator, DestAccessor> dest,
01399                                  double a0, double a1, double a2)
01400 {
01401     quadraticNoiseNormalization(src.first, src.second, src.third, dest.first, dest.second, a0, a1, a2);
01402 }
01403 
01404 /********************************************************/
01405 /*                                                      */
01406 /*                linearNoiseNormalization              */
01407 /*                                                      */
01408 /********************************************************/
01409 
01410 /** \brief Noise normalization by means of an estimated linear noise model.
01411 
01412     This function works in the same way as \ref nonparametricNoiseNormalization() with the exception of the 
01413     model for the dependency between intensity and noise variance: it assumes that this dependency is a 
01414     linear function rather than a piecewise linear function. If the linear model is applicable, it leads
01415     to a very simple transformation which is similar to the familiar gamma correction.
01416     
01417     <b> Declarations:</b>
01418     
01419     pass arguments explicitly:
01420     \code
01421     namespace vigra {
01422         template <class SrcIterator, class SrcAccessor,
01423                 class DestIterator, class DestAccessor>
01424         bool linearNoiseNormalization(SrcIterator sul, SrcIterator slr, SrcAccessor src,
01425                                       DestIterator dul, DestAccessor dest,
01426                                       NoiseNormalizationOptions const & options = NoiseNormalizationOptions());
01427     }
01428     \endcode
01429     
01430     use argument objects in conjunction with \ref ArgumentObjectFactories :
01431     \code
01432     namespace vigra {
01433         template <class SrcIterator, class SrcAccessor,
01434                   class DestIterator, class DestAccessor>
01435         bool linearNoiseNormalization(triple<SrcIterator, SrcIterator, SrcAccessor> src,
01436                                       pair<DestIterator, DestAccessor> dest,
01437                                       NoiseNormalizationOptions const & options = NoiseNormalizationOptions());
01438     }
01439     \endcode
01440     
01441     <b> Usage:</b>
01442     
01443         <b>\#include</b> <vigra/noise_normalization.hxx><br>
01444     Namespace: vigra
01445     
01446     \code
01447     vigra::BRGBImage src(w,h), dest(w, h);
01448     
01449     ...
01450     vigra::linearNoiseNormalization(srcImageRange(src), destImage(dest), 
01451                                     vigra::NoiseNormalizationOptions().windowRadius(9).noiseVarianceInitialGuess(25.0).
01452                                     clusterCount(15));
01453     \endcode
01454 
01455     <b> Required Interface:</b>
01456     
01457     same as \ref noiseVarianceEstimation()
01458 */
01459 doxygen_overloaded_function(template <...> bool linearNoiseNormalization)
01460 
01461 template <class SrcIterator, class SrcAccessor,
01462           class DestIterator, class DestAccessor>
01463 inline bool
01464 linearNoiseNormalization(SrcIterator sul, SrcIterator slr, SrcAccessor src,
01465                                 DestIterator dul, DestAccessor dest,
01466                                 NoiseNormalizationOptions const & options = NoiseNormalizationOptions())
01467 {
01468     typedef typename SrcAccessor::value_type SrcType;
01469 
01470     return detail::linearNoiseNormalizationImpl(sul, slr, src, dul, dest, options,
01471                                          typename NumericTraits<SrcType>::isScalar());
01472 }
01473 
01474 template <class SrcIterator, class SrcAccessor,
01475           class DestIterator, class DestAccessor>
01476 inline
01477 bool linearNoiseNormalization(triple<SrcIterator, SrcIterator, SrcAccessor> src,
01478                                      pair<DestIterator, DestAccessor> dest,
01479                                      NoiseNormalizationOptions const & options = NoiseNormalizationOptions())
01480 {
01481     return linearNoiseNormalization(src.first, src.second, src.third, dest.first, dest.second, options);
01482 }
01483 
01484 /********************************************************/
01485 /*                                                      */
01486 /*                 linearNoiseNormalization             */
01487 /*                                                      */
01488 /********************************************************/
01489 
01490 /** \brief Noise normalization by means of a given linear noise model.
01491 
01492     This function works similar to \ref nonparametricNoiseNormalization() with the exception that the 
01493     functional dependency of the noise variance from the intensity is given (as a linear function) 
01494     rather than estimated:
01495     
01496     \code
01497     variance = a0 + a1 * intensity
01498     \endcode
01499     
01500     <b> Declarations:</b>
01501     
01502     pass arguments explicitly:
01503     \code
01504     namespace vigra {
01505         template <class SrcIterator, class SrcAccessor,
01506                   class DestIterator, class DestAccessor>
01507         void linearNoiseNormalization(SrcIterator sul, SrcIterator slr, SrcAccessor src,
01508                                       DestIterator dul, DestAccessor dest,
01509                                       double a0, double a1);
01510     }
01511     \endcode
01512     
01513     use argument objects in conjunction with \ref ArgumentObjectFactories :
01514     \code
01515     namespace vigra {
01516         template <class SrcIterator, class SrcAccessor,
01517                   class DestIterator, class DestAccessor>
01518         void linearNoiseNormalization(triple<SrcIterator, SrcIterator, SrcAccessor> src,
01519                                       pair<DestIterator, DestAccessor> dest,
01520                                       double a0, double a1);
01521     }
01522     \endcode
01523     
01524     <b> Usage:</b>
01525     
01526         <b>\#include</b> <vigra/noise_normalization.hxx><br>
01527     Namespace: vigra
01528     
01529     \code
01530     vigra::BRGBImage src(w,h), dest(w, h);
01531     
01532     ...
01533     vigra::linearNoiseNormalization(srcImageRange(src), destImage(dest), 
01534                                     100, 0.02);
01535     \endcode
01536 
01537     <b> Required Interface:</b>
01538     
01539     The source value type must be convertible to <tt>double</tt> or must be a vector whose elements 
01540     are convertible to <tt>double</tt>. Likewise, the destination type must be assignable from <tt>double</tt>
01541     or a vector whose elements are assignable from <tt>double</tt>.
01542 */
01543 doxygen_overloaded_function(template <...> void linearNoiseNormalization)
01544 
01545 template <class SrcIterator, class SrcAccessor,
01546           class DestIterator, class DestAccessor>
01547 inline
01548 void linearNoiseNormalization(SrcIterator sul, SrcIterator slr, SrcAccessor src,
01549                               DestIterator dul, DestAccessor dest,
01550                               double a0, double a1)
01551 {
01552     typedef typename SrcAccessor::value_type SrcType;
01553 
01554     detail::linearNoiseNormalizationImpl(sul, slr, src, dul, dest, a0, a1,
01555                                          typename NumericTraits<SrcType>::isScalar());
01556 }
01557 
01558 template <class SrcIterator, class SrcAccessor,
01559           class DestIterator, class DestAccessor>
01560 inline
01561 void linearNoiseNormalization(triple<SrcIterator, SrcIterator, SrcAccessor> src,
01562                               pair<DestIterator, DestAccessor> dest,
01563                               double a0, double a1)
01564 {
01565     linearNoiseNormalization(src.first, src.second, src.third, dest.first, dest.second, a0, a1);
01566 }
01567 
01568 //@}
01569 
01570 } // namespace vigra
01571 
01572 #endif // VIGRA_NOISE_NORMALIZATION_HXX

© Ullrich Köthe (ullrich.koethe@iwr.uni-heidelberg.de)
Heidelberg Collaboratory for Image Processing, University of Heidelberg, Germany

html generated using doxygen and Python
vigra 1.8.0 (20 Sep 2011)