This function is useful for the doc. It should be used each time the output value is an image of labels (most of them are dark). Index: olena/ChangeLog from Niels Van Vliet &lt;niels(a)lrde.epita.fr&gt; * olena/oln/convert/stretch.hh: Add stretch_balance. Add comments. Index: olena/oln/convert/stretch.hh --- olena/oln/convert/stretch.hh Thu, 31 Jul 2003 18:08:39 +0200 burrus_n (oln/f/51_stretch.hh 1.13 600) +++ olena/oln/convert/stretch.hh Sat, 13 Mar 2004 17:01:12 +0100 van-vl_n (oln/f/51_stretch.hh 1.13 600) @@ -1,4 +1,4 @@ -// Copyright (C) 2001, 2002, 2003 EPITA Research and Development Laboratory +// Copyright (C) 2001, 2002, 2003, 2004 EPITA Research and Development Laboratory // // This file is part of the Olena Library. This library is free // software; you can redistribute it and/or modify it under the terms @@ -34,10 +34,35 @@ # include <oln/convert/abstract/conversion.hh> +# include <set> + namespace oln { namespace convert { + /*! Functor to stretch a value from a range Input to a range Output. + ** + ** \see stretch_balance + ** \code + ** #include <oln/basics2d.hh> + ** #include <oln/convert/stretch.hh> + ** #include <ntg/all.hh> + ** #include <iostream> + ** + ** int main() + ** { + ** oln::image2d<ntg::int_u8> lena = oln::load(IMG_IN "lena256.pgm"); + ** + ** oln::io::save(apply(oln::convert::stretch<ntg::int_u<3> >(), lena), + ** IMG_OUT "oln_convert_stretch.pgm"); + ** } + ** \endcode + ** \image html lena256.png + ** \image latex lena256.png + ** => + ** \image html oln_convert_stretch.png + ** \image latex oln_convert_stretch.png + */ template<class Output, class Exact = mlc::final> struct stretch : public abstract::conversion_to_type<Output, typename mlc::exact_vt<stretch<Output, Exact>, Exact>::ret > @@ -62,7 +87,124 @@ } }; + /*! Stretch the value of an image. + ** + ** This function stretches values between \a min_in and \a max_in + ** of an image \in, to a range that goes from \a min_out to \a max_out. + ** \arg in Input image, must be have scalar values + ** \arg min_in Lower bound of the range in the input. All values smaller + ** than min_in are converted to \a min_out. + ** \arg max_in Upper bound of the range in the input. All values + ** greater than max_in + ** are converted to \a max_out. + ** \arg min_out Low bound of the range in the output. + ** \arg max_out Upper bound of the range in the output. + ** + ** Difference between "apply(stretch<T>(), im)" and "stretch_balance<T>(im)": + ** the first one stretches all the range of oln_value_type(T), the second + ** stretches only values that appear is the image: + ** \code + ** #include <oln/basics1d.hh> + ** #include <oln/convert/stretch.hh> + ** #include <ntg/all.hh> + ** #include <iostream> + ** + ** int main() + ** { + ** + ** oln::image1d<ntg::int_u<17> > im(4); + ** im(0) = 0; im(1) = 2; im(2) = 5000; im(3) = 4000; + ** + ** // print "0 170 511 340" + ** //(values of the *image* are dispatched on the range [0..2^9-1]) + ** std::cout << oln::convert::stretch_balance<ntg::int_u<9> >(im) << std::endl; + ** + ** //print " 0 0 19 16" (19 is the result of 5000 * 2^9 / 2^17) + ** //(values of *int_u<17>* are dispatched on the range [0..2^9-1]) + ** std::cout << apply(oln::convert::stretch<ntg::int_u<9> >(), im) << std::endl; + ** } + ** \endcode + ** + ** This function is useful to stretch images of label: + ** \code + ** #include <oln/basics2d.hh> + ** #include <oln/convert/stretch.hh> + ** #include <oln/level/cc.hh> + ** #include <ntg/all.hh> + ** #include <iostream> + ** + ** int main() + ** { + ** oln::image2d<ntg::bin> light = oln::load(IMG_IN "face_se.pbm"); + ** + ** //Extraction of the connected components: + ** unsigned card; + ** oln::image2d<ntg::int_u8> cc + ** = oln::level::frontp_connected_component<ntg::int_u8>(light, + ** oln::neighb_c8(), + ** card); + ** oln::io::save(cc, "oln_convert_stretch_dark.pgm"); + ** oln::io::save(oln::convert::stretch_balance<ntg::int_u8>(cc), + ** "oln_convert_stretch_balance.pgm"); + ** } + ** \endcode + ** \image html face_se.png + ** \image latex face_se.png + ** => Without stretch_balance: + ** \image html oln_convert_stretch_dark.png + ** \image latex oln_convert_stretch_dark.png + ** => With stretch_balance: + ** \image html oln_convert_stretch_balance.png + ** \image latex oln_convert_stretch_balance.png + */ + template<class DestValue, class I> + typename mute<I, DestValue>::ret + stretch_balance(const oln::abstract::non_vectorial_image<I> &in, + const oln_value_type(I) & min_in + = ntg_min_val(oln_value_type(I)), + const oln_value_type(I) & max_in + = ntg_max_val(oln_value_type(I)), + const DestValue & min_out = ntg_min_val(DestValue), + const DestValue & max_out = ntg_max_val(DestValue)) + { + typedef typename + ntg_is_a(DestValue, ntg::non_vectorial)::ensure_type ensure_type; + + typename mute<I, DestValue>::ret out(in.size()); + //FIXME: I would like to remove the static_cast. + ntg_cumul_type(DestValue) tab[static_cast<int>(max_in - min_in + 1)]; + typedef typename std::set<oln_value_type(I)> set_type; + set_type s; + oln_iter_type(I) it(in); + + for_all(it) + if (in[it] <= max_in && in[it] >= min_in) + s.insert(in[it]); + if (s.size() == 1) + { + for_all(it) + out[it] = ntg_zero_val(DestValue); + return out; + } + { + unsigned cpt = 0; + for (typename set_type::const_iterator it(s.begin()); + it != s.end(); ++it, ++cpt) + tab[*it - min_in] = cpt * (max_out - min_out) / (s.size() - 1); + } + for_all(it) + if (min_in <= in[it]) + { + if (in[it] <= max_in) + out[it] = tab[in[it] - min_in] + min_out; + else + out[it] = max_out; + } + else + out[it] = min_out; + return out; + } } // end of convert. } // end of oln.