| Line | Branch | Exec | Source |
|---|---|---|---|
| 1 | /* | ||
| 2 | * Copyright (c) 2000-2022 Inria | ||
| 3 | * All rights reserved. | ||
| 4 | * | ||
| 5 | * Redistribution and use in source and binary forms, with or without | ||
| 6 | * modification, are permitted provided that the following conditions are met: | ||
| 7 | * | ||
| 8 | * * Redistributions of source code must retain the above copyright notice, | ||
| 9 | * this list of conditions and the following disclaimer. | ||
| 10 | * * Redistributions in binary form must reproduce the above copyright notice, | ||
| 11 | * this list of conditions and the following disclaimer in the documentation | ||
| 12 | * and/or other materials provided with the distribution. | ||
| 13 | * * Neither the name of the ALICE Project-Team nor the names of its | ||
| 14 | * contributors may be used to endorse or promote products derived from this | ||
| 15 | * software without specific prior written permission. | ||
| 16 | * | ||
| 17 | * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" | ||
| 18 | * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE | ||
| 19 | * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE | ||
| 20 | * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE | ||
| 21 | * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR | ||
| 22 | * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF | ||
| 23 | * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS | ||
| 24 | * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN | ||
| 25 | * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) | ||
| 26 | * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE | ||
| 27 | * POSSIBILITY OF SUCH DAMAGE. | ||
| 28 | * | ||
| 29 | * Contact: Bruno Levy | ||
| 30 | * | ||
| 31 | * https://www.inria.fr/fr/bruno-levy | ||
| 32 | * | ||
| 33 | * Inria, | ||
| 34 | * Domaine de Voluceau, | ||
| 35 | * 78150 Le Chesnay - Rocquencourt | ||
| 36 | * FRANCE | ||
| 37 | * | ||
| 38 | */ | ||
| 39 | |||
| 40 | #include <exploragram/optimal_transport/sampling.h> | ||
| 41 | #include <geogram/voronoi/CVT.h> | ||
| 42 | #include <geogram/mesh/mesh.h> | ||
| 43 | #include <geogram/mesh/mesh_AABB.h> | ||
| 44 | #include <geogram/mesh/mesh_tetrahedralize.h> | ||
| 45 | #include <geogram/mesh/mesh_repair.h> | ||
| 46 | #include <geogram/mesh/mesh_geometry.h> | ||
| 47 | #include <geogram/mesh/mesh_reorder.h> | ||
| 48 | #include <geogram/basic/command_line.h> | ||
| 49 | #include <geogram/basic/permutation.h> | ||
| 50 | #include <geogram/basic/progress.h> | ||
| 51 | |||
| 52 | #ifdef GEOGRAM_WITH_VORPALINE | ||
| 53 | #include <vorpalib/voronoi/LpCVT.h> | ||
| 54 | #endif | ||
| 55 | |||
| 56 | namespace { | ||
| 57 | using namespace GEO; | ||
| 58 | |||
| 59 | |||
| 60 | /** | ||
| 61 | * \brief Reorders the points in a Centroidal Voronoi Tesselation | ||
| 62 | * in such a way that continguous index ranges correspond to | ||
| 63 | * multiple resolutions. | ||
| 64 | * \param[in,out] CVT the CentroidalVoronoiTesselation | ||
| 65 | * \param[out] levels sample indices that correspond to level l are | ||
| 66 | * in the range levels[l] (included) ... levels[l+1] (excluded) | ||
| 67 | * \param[in] ratio number of samples ratio between two consecutive | ||
| 68 | * levels | ||
| 69 | * \param[in] threshold minimum number of samples in a level | ||
| 70 | */ | ||
| 71 | ✗ | void BRIO_reorder( | |
| 72 | CentroidalVoronoiTesselation& CVT, | ||
| 73 | vector<index_t>& levels, | ||
| 74 | double ratio, | ||
| 75 | index_t threshold | ||
| 76 | ) { | ||
| 77 | ✗ | vector<index_t> sorted_indices; | |
| 78 | ✗ | compute_BRIO_order( | |
| 79 | ✗ | CVT.nb_points(), CVT.embedding(0), sorted_indices, | |
| 80 | ✗ | CVT.dimension(), CVT.dimension(), threshold, ratio, &levels | |
| 81 | ); | ||
| 82 | ✗ | Permutation::apply( | |
| 83 | ✗ | CVT.embedding(0), sorted_indices, | |
| 84 | ✗ | index_t(CVT.dimension() * sizeof(double)) | |
| 85 | ); | ||
| 86 | ✗ | } | |
| 87 | |||
| 88 | /** | ||
| 89 | * \brief Internal implementation function for | ||
| 90 | * compute_hierarchical_sampling(). | ||
| 91 | * \param[in,out] CVT the CentroidalVoronoiTesselation, initialized | ||
| 92 | * with the volume to be sampled. On output, it stores the samples | ||
| 93 | * \param[in] nb_samples total number of samples to generate | ||
| 94 | * \param[out] levels sample indices that correspond to level l are | ||
| 95 | * in the range levels[l] (included) ... levels[l+1] (excluded) | ||
| 96 | * \param[in] ratio number of samples ratio between two consecutive | ||
| 97 | * levels | ||
| 98 | * \param[in] threshold minimum number of samples in a level | ||
| 99 | * \param[in] b first element of the level to be generated | ||
| 100 | * \param[in] e one position past the last element of the | ||
| 101 | * level to be generated | ||
| 102 | * \param[in,out] points work vector allocated by caller, | ||
| 103 | * of size 3*nb_samples | ||
| 104 | */ | ||
| 105 | ✗ | void compute_hierarchical_sampling_recursive( | |
| 106 | CentroidalVoronoiTesselation& CVT, | ||
| 107 | index_t nb_samples, | ||
| 108 | vector<index_t>& levels, | ||
| 109 | double ratio, | ||
| 110 | index_t threshold, | ||
| 111 | index_t b, index_t e, | ||
| 112 | vector<double>& points | ||
| 113 | ) { | ||
| 114 | ✗ | index_t m = b; | |
| 115 | |||
| 116 | // Recurse in [b...m) range | ||
| 117 | ✗ | if(e - b > threshold) { | |
| 118 | ✗ | m = b + index_t(double(e - b) * ratio); | |
| 119 | ✗ | compute_hierarchical_sampling_recursive( | |
| 120 | CVT, nb_samples, levels, ratio, threshold, b, m, points | ||
| 121 | ); | ||
| 122 | } | ||
| 123 | |||
| 124 | // Initialize random points in [m...e) range | ||
| 125 | ✗ | CVT.RVD()->compute_initial_sampling(&points[3 * m], e - m); | |
| 126 | |||
| 127 | // Set the points in [b...e) range | ||
| 128 | ✗ | CVT.set_points(e - b, &points[0]); | |
| 129 | |||
| 130 | // Lock [b...m) range | ||
| 131 | { | ||
| 132 | ✗ | for(index_t i = b; i < e; ++i) { | |
| 133 | ✗ | if(i < m) { | |
| 134 | ✗ | CVT.lock_point(i); | |
| 135 | } else { | ||
| 136 | ✗ | CVT.unlock_point(i); | |
| 137 | } | ||
| 138 | } | ||
| 139 | } | ||
| 140 | |||
| 141 | ✗ | Logger::div( | |
| 142 | ✗ | std::string("Generating level ") + | |
| 143 | ✗ | String::to_string(levels.size()) | |
| 144 | ); | ||
| 145 | |||
| 146 | ✗ | Logger::out("Sample") << " generating a level with " << e - m | |
| 147 | ✗ | << " samples" << std::endl; | |
| 148 | |||
| 149 | try { | ||
| 150 | ✗ | ProgressTask progress("Lloyd", 100); | |
| 151 | ✗ | CVT.set_progress_logger(&progress); | |
| 152 | ✗ | CVT.Lloyd_iterations(CmdLine::get_arg_uint("opt:nb_Lloyd_iter")); | |
| 153 | ✗ | } | |
| 154 | ✗ | catch(const TaskCanceled&) { | |
| 155 | ✗ | } | |
| 156 | |||
| 157 | try { | ||
| 158 | ✗ | ProgressTask progress("Newton", 100); | |
| 159 | ✗ | CVT.set_progress_logger(&progress); | |
| 160 | ✗ | CVT.Newton_iterations(CmdLine::get_arg_uint("opt:nb_Newton_iter")); | |
| 161 | ✗ | } | |
| 162 | ✗ | catch(const TaskCanceled&) { | |
| 163 | ✗ | } | |
| 164 | |||
| 165 | ✗ | levels.push_back(e); | |
| 166 | ✗ | } | |
| 167 | |||
| 168 | /** | ||
| 169 | * \brief Computes a hierarchical sampling of a volume. | ||
| 170 | * \param[in,out] CVT the CentroidalVoronoiTesselation, initialized | ||
| 171 | * with the volume to be sampled. On output, it stores the samples | ||
| 172 | * \param[in] nb_samples total number of samples to generate | ||
| 173 | * \param[out] levels sample indices that correspond to level l are | ||
| 174 | * in the range levels[l] (included) ... levels[l+1] (excluded) | ||
| 175 | * \param[in] ratio number of samples ratio between two consecutive | ||
| 176 | * levels | ||
| 177 | * \param[in] threshold minimum number of samples in a level | ||
| 178 | */ | ||
| 179 | ✗ | void compute_hierarchical_sampling( | |
| 180 | CentroidalVoronoiTesselation& CVT, | ||
| 181 | index_t nb_samples, | ||
| 182 | vector<index_t>& levels, | ||
| 183 | double ratio = 0.125, | ||
| 184 | index_t threshold = 300 | ||
| 185 | ) { | ||
| 186 | ✗ | levels.push_back(0); | |
| 187 | ✗ | vector<double> points(nb_samples * 3); | |
| 188 | ✗ | compute_hierarchical_sampling_recursive( | |
| 189 | CVT, nb_samples, levels, ratio, threshold, | ||
| 190 | 0, nb_samples, | ||
| 191 | points | ||
| 192 | ); | ||
| 193 | ✗ | CVT.unlock_all_points(); | |
| 194 | ✗ | } | |
| 195 | |||
| 196 | /** | ||
| 197 | * \brief Computes a sampling of a volume. | ||
| 198 | * \param[in,out] CVT the CentroidalVoronoiTesselation, initialized | ||
| 199 | * with the volume to be sampled. On output, it stores the samples | ||
| 200 | * \param[in] nb_samples total number of samples to generate | ||
| 201 | */ | ||
| 202 | ✗ | void compute_single_level_sampling( | |
| 203 | CentroidalVoronoiTesselation& CVT, | ||
| 204 | index_t nb_samples | ||
| 205 | ) { | ||
| 206 | |||
| 207 | ✗ | CVT.compute_initial_sampling(nb_samples); | |
| 208 | |||
| 209 | try { | ||
| 210 | ✗ | ProgressTask progress("Lloyd", 100); | |
| 211 | ✗ | CVT.set_progress_logger(&progress); | |
| 212 | ✗ | CVT.Lloyd_iterations(CmdLine::get_arg_uint("opt:nb_Lloyd_iter")); | |
| 213 | ✗ | } | |
| 214 | ✗ | catch(const TaskCanceled&) { | |
| 215 | ✗ | } | |
| 216 | |||
| 217 | try { | ||
| 218 | ✗ | ProgressTask progress("Newton", 100); | |
| 219 | ✗ | CVT.set_progress_logger(&progress); | |
| 220 | ✗ | CVT.Newton_iterations(CmdLine::get_arg_uint("opt:nb_Newton_iter")); | |
| 221 | ✗ | } | |
| 222 | ✗ | catch(const TaskCanceled&) { | |
| 223 | ✗ | } | |
| 224 | ✗ | } | |
| 225 | |||
| 226 | /** | ||
| 227 | * \brief Projects the points of a volumetric sampling | ||
| 228 | * onto the border of the volume. | ||
| 229 | */ | ||
| 230 | ✗ | void project_sampling_on_border( | |
| 231 | CentroidalVoronoiTesselation& CVT | ||
| 232 | ) { | ||
| 233 | try { | ||
| 234 | ✗ | ProgressTask progress("Surf. Lloyd", 100); | |
| 235 | ✗ | CVT.set_progress_logger(&progress); | |
| 236 | ✗ | CVT.set_volumetric(false); | |
| 237 | ✗ | CVT.Lloyd_iterations( | |
| 238 | ✗ | CmdLine::get_arg_uint("opt:nb_Lloyd_iter") * 2 | |
| 239 | ); | ||
| 240 | ✗ | } | |
| 241 | ✗ | catch(const TaskCanceled&) { | |
| 242 | ✗ | } | |
| 243 | |||
| 244 | { | ||
| 245 | |||
| 246 | #ifdef GEOGRAM_WITH_VORPALINExxx | ||
| 247 | CVT.done_current(); | ||
| 248 | { | ||
| 249 | LpCentroidalVoronoiTesselation LpCVT( | ||
| 250 | CVT.mesh(), 0 | ||
| 251 | ); | ||
| 252 | LpCVT.set_points(CVT.nb_points(), CVT.embedding(0)); | ||
| 253 | try { | ||
| 254 | ProgressTask progress("LpCVT", 100); | ||
| 255 | LpCVT.set_progress_logger(&progress); | ||
| 256 | LpCVT.set_normal_anisotropy(5.0); | ||
| 257 | LpCVT.Newton_iterations(30, 7); | ||
| 258 | } | ||
| 259 | catch(const TaskCanceled&) { | ||
| 260 | } | ||
| 261 | CVT.set_points(LpCVT.nb_points(), LpCVT.embedding(0)); | ||
| 262 | } | ||
| 263 | CVT.make_current(); | ||
| 264 | #endif | ||
| 265 | } | ||
| 266 | |||
| 267 | ✗ | vector<double> mg(3 * CVT.nb_points()); | |
| 268 | ✗ | vector<double> m(CVT.nb_points()); | |
| 269 | ✗ | CVT.RVD()->compute_centroids(&mg[0], &m[0]); | |
| 270 | ✗ | for(index_t i = 0; i < CVT.nb_points(); ++i) { | |
| 271 | ✗ | if(m[i] == 0.0) { | |
| 272 | ✗ | CVT.unlock_point(i); | |
| 273 | } else { | ||
| 274 | ✗ | CVT.lock_point(i); | |
| 275 | } | ||
| 276 | } | ||
| 277 | |||
| 278 | ✗ | CVT.set_volumetric(true); | |
| 279 | |||
| 280 | try { | ||
| 281 | ✗ | ProgressTask progress("Relax. vol.", 100); | |
| 282 | ✗ | CVT.set_progress_logger(&progress); | |
| 283 | ✗ | CVT.Lloyd_iterations( | |
| 284 | ✗ | CmdLine::get_arg_uint("opt:nb_Lloyd_iter") * 2 | |
| 285 | ); | ||
| 286 | ✗ | } | |
| 287 | ✗ | catch(const TaskCanceled&) { | |
| 288 | ✗ | } | |
| 289 | ✗ | } | |
| 290 | |||
| 291 | } | ||
| 292 | |||
| 293 | namespace GEO { | ||
| 294 | |||
| 295 | ✗ | void recenter_mesh(const Mesh& M1, Mesh& M2) { | |
| 296 | double xyzmin1[3]; | ||
| 297 | double xyzmax1[3]; | ||
| 298 | double xyzmin2[3]; | ||
| 299 | double xyzmax2[3]; | ||
| 300 | double xlat[3]; | ||
| 301 | ✗ | get_bbox(M1, xyzmin1, xyzmax1); | |
| 302 | ✗ | get_bbox(M2, xyzmin2, xyzmax2); | |
| 303 | ✗ | for(coord_index_t c=0; c<3; ++c) { | |
| 304 | ✗ | xlat[c] = 0.5* | |
| 305 | ✗ | ((xyzmin1[c] + xyzmax1[c]) - (xyzmin2[c] + xyzmax2[c])); | |
| 306 | } | ||
| 307 | ✗ | for(index_t v=0; v<M2.vertices.nb(); ++v) { | |
| 308 | ✗ | for(coord_index_t c=0; c<3; ++c) { | |
| 309 | ✗ | M2.vertices.point_ptr(v)[c] += xlat[c]; | |
| 310 | } | ||
| 311 | } | ||
| 312 | ✗ | } | |
| 313 | |||
| 314 | ✗ | double mesh_tets_volume(const Mesh& M) { | |
| 315 | ✗ | double result = 0.0; | |
| 316 | ✗ | for(index_t t = 0; t < M.cells.nb(); ++t) { | |
| 317 | ✗ | result += Geom::tetra_volume<3>( | |
| 318 | M.vertices.point_ptr(M.cells.tet_vertex(t, 0)), | ||
| 319 | M.vertices.point_ptr(M.cells.tet_vertex(t, 1)), | ||
| 320 | M.vertices.point_ptr(M.cells.tet_vertex(t, 2)), | ||
| 321 | M.vertices.point_ptr(M.cells.tet_vertex(t, 3)) | ||
| 322 | ); | ||
| 323 | } | ||
| 324 | ✗ | return result; | |
| 325 | } | ||
| 326 | |||
| 327 | ✗ | void rescale_mesh(const Mesh& M1, Mesh& M2) { | |
| 328 | double xyzmin[3]; | ||
| 329 | double xyzmax[3]; | ||
| 330 | ✗ | get_bbox(M2, xyzmin, xyzmax); | |
| 331 | ✗ | double s = pow(mesh_tets_volume(M1)/mesh_tets_volume(M2), 1.0/3.0); | |
| 332 | ✗ | for(unsigned int v=0; v<M2.vertices.nb(); ++v) { | |
| 333 | ✗ | for(index_t c=0; c<3; ++c) { | |
| 334 | ✗ | double gc = 0.5*(xyzmin[c]+xyzmax[c]); | |
| 335 | ✗ | M2.vertices.point_ptr(v)[c] = | |
| 336 | ✗ | gc + s * (M2.vertices.point_ptr(v)[c] - gc); | |
| 337 | } | ||
| 338 | } | ||
| 339 | ✗ | } | |
| 340 | |||
| 341 | enum DensityFunction { | ||
| 342 | DENSITY_X=0, | ||
| 343 | DENSITY_Y=1, | ||
| 344 | DENSITY_Z=2, | ||
| 345 | DENSITY_R, | ||
| 346 | DENSITY_SIN, | ||
| 347 | DENSITY_DIST | ||
| 348 | }; | ||
| 349 | |||
| 350 | ✗ | void set_density( | |
| 351 | Mesh& M, double mass1, double mass2, const std::string& function_str_in, | ||
| 352 | Mesh* density_distance_reference | ||
| 353 | ) { | ||
| 354 | ✗ | std::string function_str = function_str_in; | |
| 355 | |||
| 356 | ✗ | if(mass1 == mass2) { | |
| 357 | ✗ | return; | |
| 358 | } | ||
| 359 | |||
| 360 | ✗ | bool minus = false; | |
| 361 | ✗ | if(function_str.length() > 1 && function_str[0] == '-') { | |
| 362 | ✗ | minus = true; | |
| 363 | ✗ | function_str = function_str.substr(1,function_str.length()-1); | |
| 364 | } | ||
| 365 | ✗ | double density_pow = 1.0; | |
| 366 | { | ||
| 367 | ✗ | std::size_t found = function_str.find('^'); | |
| 368 | ✗ | if(found != std::string::npos) { | |
| 369 | std::string pow_str = | ||
| 370 | ✗ | function_str.substr(found+1, function_str.length()-found-1); | |
| 371 | ✗ | density_pow = String::to_double(pow_str); | |
| 372 | ✗ | function_str = function_str.substr(0,found); | |
| 373 | ✗ | } | |
| 374 | } | ||
| 375 | |||
| 376 | ✗ | Logger::out("OTM") | |
| 377 | << "Using density: " | ||
| 378 | << (minus ? "-" : "+") | ||
| 379 | ✗ | << function_str << "^" | |
| 380 | ✗ | << density_pow | |
| 381 | ✗ | << " rescaled to (" | |
| 382 | ✗ | << mass1 << "," << mass2 | |
| 383 | ✗ | << ")" | |
| 384 | ✗ | << std::endl; | |
| 385 | |||
| 386 | DensityFunction function; | ||
| 387 | ✗ | if(function_str == "X") { | |
| 388 | ✗ | function = DENSITY_X; | |
| 389 | ✗ | } else if(function_str == "Y") { | |
| 390 | ✗ | function = DENSITY_Y; | |
| 391 | ✗ | } else if(function_str == "Z") { | |
| 392 | ✗ | function = DENSITY_Z; | |
| 393 | ✗ | } else if(function_str == "R") { | |
| 394 | ✗ | function = DENSITY_R; | |
| 395 | ✗ | } else if(function_str == "sin") { | |
| 396 | ✗ | function = DENSITY_SIN; | |
| 397 | ✗ | } else if(function_str == "dist") { | |
| 398 | ✗ | function = DENSITY_DIST; | |
| 399 | } else { | ||
| 400 | ✗ | Logger::err("OTM") << function_str << ": no such density function" | |
| 401 | ✗ | << std::endl; | |
| 402 | ✗ | return; | |
| 403 | } | ||
| 404 | |||
| 405 | ✗ | Attribute<double> mass(M.vertices.attributes(),"weight"); | |
| 406 | |||
| 407 | ✗ | switch(function) { | |
| 408 | ✗ | case DENSITY_X: | |
| 409 | case DENSITY_Y: | ||
| 410 | case DENSITY_Z: { | ||
| 411 | ✗ | for(index_t v=0; v<M.vertices.nb(); ++v) { | |
| 412 | ✗ | mass[v] = M.vertices.point_ptr(v)[index_t(function)]; | |
| 413 | } | ||
| 414 | ✗ | } break; | |
| 415 | ✗ | case DENSITY_R: { | |
| 416 | double xyz_min[3]; | ||
| 417 | double xyz_max[3]; | ||
| 418 | ✗ | get_bbox(M, xyz_min, xyz_max); | |
| 419 | ✗ | for(index_t v=0; v<M.vertices.nb(); ++v) { | |
| 420 | ✗ | double r=0; | |
| 421 | ✗ | const double* p = M.vertices.point_ptr(v); | |
| 422 | ✗ | for(coord_index_t c=0; c<3; ++c) { | |
| 423 | ✗ | r += geo_sqr(p[c] - 0.5*(xyz_min[c] + xyz_max[c])); | |
| 424 | } | ||
| 425 | ✗ | r = ::sqrt(r); | |
| 426 | ✗ | mass[v] = r; | |
| 427 | } | ||
| 428 | ✗ | } break; | |
| 429 | ✗ | case DENSITY_SIN: { | |
| 430 | double xyz_min[3]; | ||
| 431 | double xyz_max[3]; | ||
| 432 | ✗ | get_bbox(M, xyz_min, xyz_max); | |
| 433 | ✗ | for(index_t v=0; v<M.vertices.nb(); ++v) { | |
| 434 | ✗ | double f = 1.0; | |
| 435 | ✗ | const double* p = M.vertices.point_ptr(v); | |
| 436 | ✗ | for(coord_index_t c=0; c<3; ++c) { | |
| 437 | ✗ | double coord = | |
| 438 | ✗ | (p[c] - xyz_min[c]) / (xyz_max[c] - xyz_min[c]); | |
| 439 | ✗ | f *= sin(coord * M_PI * 2.0 * 2.0); | |
| 440 | } | ||
| 441 | ✗ | mass[v] = f; | |
| 442 | } | ||
| 443 | ✗ | } break; | |
| 444 | ✗ | case DENSITY_DIST: { | |
| 445 | ✗ | if(density_distance_reference != nullptr) { | |
| 446 | ✗ | MeshFacetsAABB AABB(*density_distance_reference); | |
| 447 | ✗ | for(index_t v=0; v<M.vertices.nb(); ++v) { | |
| 448 | ✗ | mass[v] = | |
| 449 | ✗ | ::sqrt(AABB.squared_distance( | |
| 450 | ✗ | vec3(M.vertices.point_ptr(v))) | |
| 451 | ); | ||
| 452 | } | ||
| 453 | ✗ | } else { | |
| 454 | ✗ | MeshFacetsAABB AABB(M); | |
| 455 | ✗ | for(index_t v=0; v<M.vertices.nb(); ++v) { | |
| 456 | ✗ | mass[v] = ::sqrt( | |
| 457 | ✗ | AABB.squared_distance(vec3(M.vertices.point_ptr(v))) | |
| 458 | ); | ||
| 459 | } | ||
| 460 | ✗ | } | |
| 461 | ✗ | } break; | |
| 462 | } | ||
| 463 | |||
| 464 | // Compute min and max mass | ||
| 465 | ✗ | double mass_min = Numeric::max_float64(); | |
| 466 | ✗ | double mass_max = Numeric::min_float64(); | |
| 467 | ✗ | for(index_t v=0; v<M.vertices.nb(); ++v) { | |
| 468 | ✗ | mass_min = std::min(mass_min, mass[v]); | |
| 469 | ✗ | mass_max = std::max(mass_max, mass[v]); | |
| 470 | } | ||
| 471 | |||
| 472 | // Normalize mass, apply power, and rescale to (mass1 - mass2) | ||
| 473 | ✗ | for(index_t v=0; v<M.vertices.nb(); ++v) { | |
| 474 | ✗ | double f = (mass[v] - mass_min) / (mass_max - mass_min); | |
| 475 | ✗ | if(minus) { | |
| 476 | ✗ | f = 1.0 - f; | |
| 477 | } | ||
| 478 | ✗ | f = ::pow(f,density_pow); | |
| 479 | ✗ | mass[v] = mass1 + f*(mass2 - mass1); | |
| 480 | } | ||
| 481 | ✗ | } | |
| 482 | |||
| 483 | ✗ | void sample( | |
| 484 | CentroidalVoronoiTesselation& CVT, | ||
| 485 | index_t nb_points, bool project_on_border, | ||
| 486 | bool BRIO, bool multilevel, double ratio, | ||
| 487 | vector<index_t>* levels_out | ||
| 488 | ) { | ||
| 489 | ✗ | vector<index_t> levels; | |
| 490 | ✗ | multilevel = multilevel | BRIO; | |
| 491 | ✗ | if(CmdLine::get_arg_bool("RVD_iter") && multilevel) { | |
| 492 | ✗ | Logger::warn("OTM") << "Deactivating multilevel mode" << std::endl; | |
| 493 | ✗ | Logger::warn("OTM") << "(because RVD_iter is set)" << std::endl; | |
| 494 | ✗ | multilevel = false; | |
| 495 | } | ||
| 496 | ✗ | if(multilevel) { | |
| 497 | ✗ | if(BRIO) { | |
| 498 | ✗ | compute_single_level_sampling(CVT, nb_points); | |
| 499 | ✗ | BRIO_reorder(CVT, levels, ratio, 300); | |
| 500 | } else { | ||
| 501 | ✗ | compute_hierarchical_sampling( | |
| 502 | CVT, nb_points,levels,ratio | ||
| 503 | ); | ||
| 504 | } | ||
| 505 | } else { | ||
| 506 | ✗ | compute_single_level_sampling(CVT, nb_points); | |
| 507 | } | ||
| 508 | ✗ | if(levels_out != nullptr) { | |
| 509 | ✗ | *levels_out = levels; | |
| 510 | } | ||
| 511 | ✗ | if(project_on_border) { | |
| 512 | ✗ | project_sampling_on_border(CVT); | |
| 513 | } | ||
| 514 | ✗ | } | |
| 515 | |||
| 516 | |||
| 517 | } | ||
| 518 |