blob: 42e8a6ed67da36c6d762a322a45822b8db68e690 [file] [log] [blame]
Austin Schuh70cc9552019-01-21 19:46:48 -08001// Ceres Solver - A fast non-linear least squares minimizer
Austin Schuh3de38b02024-06-25 18:25:10 -07002// Copyright 2023 Google Inc. All rights reserved.
Austin Schuh70cc9552019-01-21 19:46:48 -08003// http://ceres-solver.org/
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 Google Inc. nor the names of its contributors may be
14// used to endorse or promote products derived from this software without
15// 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 OWNER 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// Author: sameeragarwal@google.com (Sameer Agarwal)
30
31#include "ceres/visibility_based_preconditioner.h"
32
33#include <algorithm>
34#include <functional>
35#include <iterator>
36#include <memory>
37#include <set>
Austin Schuh3de38b02024-06-25 18:25:10 -070038#include <string>
39#include <unordered_set>
Austin Schuh70cc9552019-01-21 19:46:48 -080040#include <utility>
41#include <vector>
42
43#include "Eigen/Dense"
44#include "ceres/block_random_access_sparse_matrix.h"
45#include "ceres/block_sparse_matrix.h"
46#include "ceres/canonical_views_clustering.h"
47#include "ceres/graph.h"
48#include "ceres/graph_algorithms.h"
49#include "ceres/linear_solver.h"
50#include "ceres/schur_eliminator.h"
51#include "ceres/single_linkage_clustering.h"
52#include "ceres/visibility.h"
53#include "glog/logging.h"
54
Austin Schuh3de38b02024-06-25 18:25:10 -070055namespace ceres::internal {
Austin Schuh70cc9552019-01-21 19:46:48 -080056
57// TODO(sameeragarwal): Currently these are magic weights for the
58// preconditioner construction. Move these higher up into the Options
59// struct and provide some guidelines for choosing them.
60//
61// This will require some more work on the clustering algorithm and
62// possibly some more refactoring of the code.
Austin Schuh1d1e6ea2020-12-23 21:56:30 -080063static constexpr double kCanonicalViewsSizePenaltyWeight = 3.0;
64static constexpr double kCanonicalViewsSimilarityPenaltyWeight = 0.0;
65static constexpr double kSingleLinkageMinSimilarity = 0.9;
Austin Schuh70cc9552019-01-21 19:46:48 -080066
67VisibilityBasedPreconditioner::VisibilityBasedPreconditioner(
Austin Schuh3de38b02024-06-25 18:25:10 -070068 const CompressedRowBlockStructure& bs, Preconditioner::Options options)
69 : options_(std::move(options)), num_blocks_(0), num_clusters_(0) {
Austin Schuh70cc9552019-01-21 19:46:48 -080070 CHECK_GT(options_.elimination_groups.size(), 1);
71 CHECK_GT(options_.elimination_groups[0], 0);
72 CHECK(options_.type == CLUSTER_JACOBI || options_.type == CLUSTER_TRIDIAGONAL)
73 << "Unknown preconditioner type: " << options_.type;
74 num_blocks_ = bs.cols.size() - options_.elimination_groups[0];
75 CHECK_GT(num_blocks_, 0) << "Jacobian should have at least 1 f_block for "
76 << "visibility based preconditioning.";
Austin Schuh3de38b02024-06-25 18:25:10 -070077 CHECK(options_.context != nullptr);
Austin Schuh70cc9552019-01-21 19:46:48 -080078
79 // Vector of camera block sizes
Austin Schuh3de38b02024-06-25 18:25:10 -070080 blocks_ = Tail(bs.cols, bs.cols.size() - options_.elimination_groups[0]);
Austin Schuh70cc9552019-01-21 19:46:48 -080081
Austin Schuh3de38b02024-06-25 18:25:10 -070082 const time_t start_time = time(nullptr);
Austin Schuh70cc9552019-01-21 19:46:48 -080083 switch (options_.type) {
84 case CLUSTER_JACOBI:
85 ComputeClusterJacobiSparsity(bs);
86 break;
87 case CLUSTER_TRIDIAGONAL:
88 ComputeClusterTridiagonalSparsity(bs);
89 break;
90 default:
91 LOG(FATAL) << "Unknown preconditioner type";
92 }
Austin Schuh3de38b02024-06-25 18:25:10 -070093 const time_t structure_time = time(nullptr);
Austin Schuh70cc9552019-01-21 19:46:48 -080094 InitStorage(bs);
Austin Schuh3de38b02024-06-25 18:25:10 -070095 const time_t storage_time = time(nullptr);
Austin Schuh70cc9552019-01-21 19:46:48 -080096 InitEliminator(bs);
Austin Schuh3de38b02024-06-25 18:25:10 -070097 const time_t eliminator_time = time(nullptr);
Austin Schuh70cc9552019-01-21 19:46:48 -080098
99 LinearSolver::Options sparse_cholesky_options;
100 sparse_cholesky_options.sparse_linear_algebra_library_type =
101 options_.sparse_linear_algebra_library_type;
Austin Schuh3de38b02024-06-25 18:25:10 -0700102 sparse_cholesky_options.ordering_type = options_.ordering_type;
Austin Schuh70cc9552019-01-21 19:46:48 -0800103 sparse_cholesky_ = SparseCholesky::Create(sparse_cholesky_options);
104
Austin Schuh3de38b02024-06-25 18:25:10 -0700105 const time_t init_time = time(nullptr);
Austin Schuh70cc9552019-01-21 19:46:48 -0800106 VLOG(2) << "init time: " << init_time - start_time
107 << " structure time: " << structure_time - start_time
108 << " storage time:" << storage_time - structure_time
109 << " eliminator time: " << eliminator_time - storage_time;
110}
111
Austin Schuh3de38b02024-06-25 18:25:10 -0700112VisibilityBasedPreconditioner::~VisibilityBasedPreconditioner() = default;
Austin Schuh70cc9552019-01-21 19:46:48 -0800113
114// Determine the sparsity structure of the CLUSTER_JACOBI
115// preconditioner. It clusters cameras using their scene
116// visibility. The clusters form the diagonal blocks of the
117// preconditioner matrix.
118void VisibilityBasedPreconditioner::ComputeClusterJacobiSparsity(
119 const CompressedRowBlockStructure& bs) {
Austin Schuh3de38b02024-06-25 18:25:10 -0700120 std::vector<std::set<int>> visibility;
Austin Schuh70cc9552019-01-21 19:46:48 -0800121 ComputeVisibility(bs, options_.elimination_groups[0], &visibility);
122 CHECK_EQ(num_blocks_, visibility.size());
123 ClusterCameras(visibility);
124 cluster_pairs_.clear();
125 for (int i = 0; i < num_clusters_; ++i) {
Austin Schuh3de38b02024-06-25 18:25:10 -0700126 cluster_pairs_.insert(std::make_pair(i, i));
Austin Schuh70cc9552019-01-21 19:46:48 -0800127 }
128}
129
130// Determine the sparsity structure of the CLUSTER_TRIDIAGONAL
Austin Schuh1d1e6ea2020-12-23 21:56:30 -0800131// preconditioner. It clusters cameras using the scene visibility and
132// then finds the strongly interacting pairs of clusters by
133// constructing another graph with the clusters as vertices and
134// approximating it with a degree-2 maximum spanning forest. The set
135// of edges in this forest are the cluster pairs.
Austin Schuh70cc9552019-01-21 19:46:48 -0800136void VisibilityBasedPreconditioner::ComputeClusterTridiagonalSparsity(
137 const CompressedRowBlockStructure& bs) {
Austin Schuh3de38b02024-06-25 18:25:10 -0700138 std::vector<std::set<int>> visibility;
Austin Schuh70cc9552019-01-21 19:46:48 -0800139 ComputeVisibility(bs, options_.elimination_groups[0], &visibility);
140 CHECK_EQ(num_blocks_, visibility.size());
141 ClusterCameras(visibility);
142
143 // Construct a weighted graph on the set of clusters, where the
144 // edges are the number of 3D points/e_blocks visible in both the
145 // clusters at the ends of the edge. Return an approximate degree-2
146 // maximum spanning forest of this graph.
Austin Schuh3de38b02024-06-25 18:25:10 -0700147 std::vector<std::set<int>> cluster_visibility;
Austin Schuh70cc9552019-01-21 19:46:48 -0800148 ComputeClusterVisibility(visibility, &cluster_visibility);
Austin Schuh3de38b02024-06-25 18:25:10 -0700149 auto cluster_graph = CreateClusterGraph(cluster_visibility);
Austin Schuh70cc9552019-01-21 19:46:48 -0800150 CHECK(cluster_graph != nullptr);
Austin Schuh3de38b02024-06-25 18:25:10 -0700151 auto forest = Degree2MaximumSpanningForest(*cluster_graph);
Austin Schuh70cc9552019-01-21 19:46:48 -0800152 CHECK(forest != nullptr);
153 ForestToClusterPairs(*forest, &cluster_pairs_);
154}
155
156// Allocate storage for the preconditioner matrix.
157void VisibilityBasedPreconditioner::InitStorage(
158 const CompressedRowBlockStructure& bs) {
159 ComputeBlockPairsInPreconditioner(bs);
Austin Schuh3de38b02024-06-25 18:25:10 -0700160 m_ = std::make_unique<BlockRandomAccessSparseMatrix>(
161 blocks_, block_pairs_, options_.context, options_.num_threads);
Austin Schuh70cc9552019-01-21 19:46:48 -0800162}
163
164// Call the canonical views algorithm and cluster the cameras based on
165// their visibility sets. The visibility set of a camera is the set of
166// e_blocks/3D points in the scene that are seen by it.
167//
168// The cluster_membership_ vector is updated to indicate cluster
169// memberships for each camera block.
170void VisibilityBasedPreconditioner::ClusterCameras(
Austin Schuh3de38b02024-06-25 18:25:10 -0700171 const std::vector<std::set<int>>& visibility) {
172 auto schur_complement_graph = CreateSchurComplementGraph(visibility);
Austin Schuh70cc9552019-01-21 19:46:48 -0800173 CHECK(schur_complement_graph != nullptr);
174
175 std::unordered_map<int, int> membership;
176
177 if (options_.visibility_clustering_type == CANONICAL_VIEWS) {
Austin Schuh3de38b02024-06-25 18:25:10 -0700178 std::vector<int> centers;
Austin Schuh70cc9552019-01-21 19:46:48 -0800179 CanonicalViewsClusteringOptions clustering_options;
180 clustering_options.size_penalty_weight = kCanonicalViewsSizePenaltyWeight;
181 clustering_options.similarity_penalty_weight =
182 kCanonicalViewsSimilarityPenaltyWeight;
183 ComputeCanonicalViewsClustering(
184 clustering_options, *schur_complement_graph, &centers, &membership);
185 num_clusters_ = centers.size();
186 } else if (options_.visibility_clustering_type == SINGLE_LINKAGE) {
187 SingleLinkageClusteringOptions clustering_options;
188 clustering_options.min_similarity = kSingleLinkageMinSimilarity;
189 num_clusters_ = ComputeSingleLinkageClustering(
190 clustering_options, *schur_complement_graph, &membership);
191 } else {
192 LOG(FATAL) << "Unknown visibility clustering algorithm.";
193 }
194
195 CHECK_GT(num_clusters_, 0);
196 VLOG(2) << "num_clusters: " << num_clusters_;
197 FlattenMembershipMap(membership, &cluster_membership_);
198}
199
200// Compute the block sparsity structure of the Schur complement
201// matrix. For each pair of cameras contributing a non-zero cell to
202// the schur complement, determine if that cell is present in the
203// preconditioner or not.
204//
205// A pair of cameras contribute a cell to the preconditioner if they
206// are part of the same cluster or if the two clusters that they
207// belong have an edge connecting them in the degree-2 maximum
208// spanning forest.
209//
210// For example, a camera pair (i,j) where i belongs to cluster1 and
211// j belongs to cluster2 (assume that cluster1 < cluster2).
212//
213// The cell corresponding to (i,j) is present in the preconditioner
214// if cluster1 == cluster2 or the pair (cluster1, cluster2) were
215// connected by an edge in the degree-2 maximum spanning forest.
216//
217// Since we have already expanded the forest into a set of camera
218// pairs/edges, including self edges, the check can be reduced to
219// checking membership of (cluster1, cluster2) in cluster_pairs_.
220void VisibilityBasedPreconditioner::ComputeBlockPairsInPreconditioner(
221 const CompressedRowBlockStructure& bs) {
222 block_pairs_.clear();
223 for (int i = 0; i < num_blocks_; ++i) {
Austin Schuh3de38b02024-06-25 18:25:10 -0700224 block_pairs_.insert(std::make_pair(i, i));
Austin Schuh70cc9552019-01-21 19:46:48 -0800225 }
226
227 int r = 0;
228 const int num_row_blocks = bs.rows.size();
229 const int num_eliminate_blocks = options_.elimination_groups[0];
230
231 // Iterate over each row of the matrix. The block structure of the
232 // matrix is assumed to be sorted in order of the e_blocks/point
233 // blocks. Thus all row blocks containing an e_block/point occur
234 // contiguously. Further, if present, an e_block is always the first
235 // parameter block in each row block. These structural assumptions
236 // are common to all Schur complement based solvers in Ceres.
237 //
238 // For each e_block/point block we identify the set of cameras
239 // seeing it. The cross product of this set with itself is the set
240 // of non-zero cells contributed by this e_block.
241 //
242 // The time complexity of this is O(nm^2) where, n is the number of
243 // 3d points and m is the maximum number of cameras seeing any
244 // point, which for most scenes is a fairly small number.
245 while (r < num_row_blocks) {
246 int e_block_id = bs.rows[r].cells.front().block_id;
247 if (e_block_id >= num_eliminate_blocks) {
248 // Skip the rows whose first block is an f_block.
249 break;
250 }
251
Austin Schuh3de38b02024-06-25 18:25:10 -0700252 std::set<int> f_blocks;
Austin Schuh70cc9552019-01-21 19:46:48 -0800253 for (; r < num_row_blocks; ++r) {
254 const CompressedRow& row = bs.rows[r];
255 if (row.cells.front().block_id != e_block_id) {
256 break;
257 }
258
259 // Iterate over the blocks in the row, ignoring the first block
260 // since it is the one to be eliminated and adding the rest to
261 // the list of f_blocks associated with this e_block.
262 for (int c = 1; c < row.cells.size(); ++c) {
263 const Cell& cell = row.cells[c];
264 const int f_block_id = cell.block_id - num_eliminate_blocks;
265 CHECK_GE(f_block_id, 0);
266 f_blocks.insert(f_block_id);
267 }
268 }
269
Austin Schuh3de38b02024-06-25 18:25:10 -0700270 for (auto block1 = f_blocks.begin(); block1 != f_blocks.end(); ++block1) {
271 auto block2 = block1;
Austin Schuh70cc9552019-01-21 19:46:48 -0800272 ++block2;
273 for (; block2 != f_blocks.end(); ++block2) {
274 if (IsBlockPairInPreconditioner(*block1, *block2)) {
Austin Schuh3de38b02024-06-25 18:25:10 -0700275 block_pairs_.emplace(*block1, *block2);
Austin Schuh70cc9552019-01-21 19:46:48 -0800276 }
277 }
278 }
279 }
280
281 // The remaining rows which do not contain any e_blocks.
282 for (; r < num_row_blocks; ++r) {
283 const CompressedRow& row = bs.rows[r];
284 CHECK_GE(row.cells.front().block_id, num_eliminate_blocks);
285 for (int i = 0; i < row.cells.size(); ++i) {
286 const int block1 = row.cells[i].block_id - num_eliminate_blocks;
Austin Schuh3de38b02024-06-25 18:25:10 -0700287 for (const auto& cell : row.cells) {
288 const int block2 = cell.block_id - num_eliminate_blocks;
Austin Schuh70cc9552019-01-21 19:46:48 -0800289 if (block1 <= block2) {
290 if (IsBlockPairInPreconditioner(block1, block2)) {
Austin Schuh3de38b02024-06-25 18:25:10 -0700291 block_pairs_.insert(std::make_pair(block1, block2));
Austin Schuh70cc9552019-01-21 19:46:48 -0800292 }
293 }
294 }
295 }
296 }
297
298 VLOG(1) << "Block pair stats: " << block_pairs_.size();
299}
300
301// Initialize the SchurEliminator.
302void VisibilityBasedPreconditioner::InitEliminator(
303 const CompressedRowBlockStructure& bs) {
304 LinearSolver::Options eliminator_options;
305 eliminator_options.elimination_groups = options_.elimination_groups;
306 eliminator_options.num_threads = options_.num_threads;
307 eliminator_options.e_block_size = options_.e_block_size;
308 eliminator_options.f_block_size = options_.f_block_size;
309 eliminator_options.row_block_size = options_.row_block_size;
310 eliminator_options.context = options_.context;
Austin Schuh3de38b02024-06-25 18:25:10 -0700311 eliminator_ = SchurEliminatorBase::Create(eliminator_options);
Austin Schuh70cc9552019-01-21 19:46:48 -0800312 const bool kFullRankETE = true;
313 eliminator_->Init(
314 eliminator_options.elimination_groups[0], kFullRankETE, &bs);
315}
316
317// Update the values of the preconditioner matrix and factorize it.
318bool VisibilityBasedPreconditioner::UpdateImpl(const BlockSparseMatrix& A,
319 const double* D) {
Austin Schuh3de38b02024-06-25 18:25:10 -0700320 const time_t start_time = time(nullptr);
Austin Schuh70cc9552019-01-21 19:46:48 -0800321 const int num_rows = m_->num_rows();
322 CHECK_GT(num_rows, 0);
323
324 // Compute a subset of the entries of the Schur complement.
Austin Schuh1d1e6ea2020-12-23 21:56:30 -0800325 eliminator_->Eliminate(
326 BlockSparseMatrixData(A), nullptr, D, m_.get(), nullptr);
Austin Schuh70cc9552019-01-21 19:46:48 -0800327
328 // Try factorizing the matrix. For CLUSTER_JACOBI, this should
329 // always succeed modulo some numerical/conditioning problems. For
330 // CLUSTER_TRIDIAGONAL, in general the preconditioner matrix as
331 // constructed is not positive definite. However, we will go ahead
332 // and try factorizing it. If it works, great, otherwise we scale
333 // all the cells in the preconditioner corresponding to the edges in
334 // the degree-2 forest and that guarantees positive
335 // definiteness. The proof of this fact can be found in Lemma 1 in
336 // "Visibility Based Preconditioning for Bundle Adjustment".
337 //
338 // Doing the factorization like this saves us matrix mass when
339 // scaling is not needed, which is quite often in our experience.
340 LinearSolverTerminationType status = Factorize();
341
Austin Schuh3de38b02024-06-25 18:25:10 -0700342 if (status == LinearSolverTerminationType::FATAL_ERROR) {
Austin Schuh70cc9552019-01-21 19:46:48 -0800343 return false;
344 }
345
346 // The scaling only affects the tri-diagonal case, since
347 // ScaleOffDiagonalBlocks only pays attention to the cells that
348 // belong to the edges of the degree-2 forest. In the CLUSTER_JACOBI
349 // case, the preconditioner is guaranteed to be positive
350 // semidefinite.
Austin Schuh3de38b02024-06-25 18:25:10 -0700351 if (status == LinearSolverTerminationType::FAILURE &&
352 options_.type == CLUSTER_TRIDIAGONAL) {
Austin Schuh70cc9552019-01-21 19:46:48 -0800353 VLOG(1) << "Unscaled factorization failed. Retrying with off-diagonal "
354 << "scaling";
355 ScaleOffDiagonalCells();
356 status = Factorize();
357 }
358
Austin Schuh3de38b02024-06-25 18:25:10 -0700359 VLOG(2) << "Compute time: " << time(nullptr) - start_time;
360 return (status == LinearSolverTerminationType::SUCCESS);
Austin Schuh70cc9552019-01-21 19:46:48 -0800361}
362
363// Consider the preconditioner matrix as meta-block matrix, whose
364// blocks correspond to the clusters. Then cluster pairs corresponding
365// to edges in the degree-2 forest are off diagonal entries of this
366// matrix. Scaling these off-diagonal entries by 1/2 forces this
367// matrix to be positive definite.
368void VisibilityBasedPreconditioner::ScaleOffDiagonalCells() {
369 for (const auto& block_pair : block_pairs_) {
370 const int block1 = block_pair.first;
371 const int block2 = block_pair.second;
372 if (!IsBlockPairOffDiagonal(block1, block2)) {
373 continue;
374 }
375
376 int r, c, row_stride, col_stride;
377 CellInfo* cell_info =
378 m_->GetCell(block1, block2, &r, &c, &row_stride, &col_stride);
Austin Schuh3de38b02024-06-25 18:25:10 -0700379 CHECK(cell_info != nullptr)
Austin Schuh70cc9552019-01-21 19:46:48 -0800380 << "Cell missing for block pair (" << block1 << "," << block2 << ")"
381 << " cluster pair (" << cluster_membership_[block1] << " "
382 << cluster_membership_[block2] << ")";
383
384 // Ah the magic of tri-diagonal matrices and diagonal
385 // dominance. See Lemma 1 in "Visibility Based Preconditioning
386 // For Bundle Adjustment".
387 MatrixRef m(cell_info->values, row_stride, col_stride);
Austin Schuh3de38b02024-06-25 18:25:10 -0700388 m.block(r, c, blocks_[block1].size, blocks_[block2].size) *= 0.5;
Austin Schuh70cc9552019-01-21 19:46:48 -0800389 }
390}
391
392// Compute the sparse Cholesky factorization of the preconditioner
393// matrix.
394LinearSolverTerminationType VisibilityBasedPreconditioner::Factorize() {
Austin Schuh3de38b02024-06-25 18:25:10 -0700395 // Extract the BlockSparseMatrix that is used for actually storing
Austin Schuh70cc9552019-01-21 19:46:48 -0800396 // S and convert it into a CompressedRowSparseMatrix.
Austin Schuh3de38b02024-06-25 18:25:10 -0700397 const BlockSparseMatrix* bsm =
398 down_cast<BlockRandomAccessSparseMatrix*>(m_.get())->matrix();
Austin Schuh70cc9552019-01-21 19:46:48 -0800399 const CompressedRowSparseMatrix::StorageType storage_type =
400 sparse_cholesky_->StorageType();
Austin Schuh3de38b02024-06-25 18:25:10 -0700401 if (storage_type ==
402 CompressedRowSparseMatrix::StorageType::UPPER_TRIANGULAR) {
403 if (!m_crs_) {
404 m_crs_ = bsm->ToCompressedRowSparseMatrix();
405 m_crs_->set_storage_type(
406 CompressedRowSparseMatrix::StorageType::UPPER_TRIANGULAR);
407 } else {
408 bsm->UpdateCompressedRowSparseMatrix(m_crs_.get());
409 }
Austin Schuh70cc9552019-01-21 19:46:48 -0800410 } else {
Austin Schuh3de38b02024-06-25 18:25:10 -0700411 if (!m_crs_) {
412 m_crs_ = bsm->ToCompressedRowSparseMatrixTranspose();
413 m_crs_->set_storage_type(
414 CompressedRowSparseMatrix::StorageType::LOWER_TRIANGULAR);
415 } else {
416 bsm->UpdateCompressedRowSparseMatrixTranspose(m_crs_.get());
417 }
Austin Schuh70cc9552019-01-21 19:46:48 -0800418 }
419
420 std::string message;
Austin Schuh3de38b02024-06-25 18:25:10 -0700421 return sparse_cholesky_->Factorize(m_crs_.get(), &message);
Austin Schuh70cc9552019-01-21 19:46:48 -0800422}
423
Austin Schuh3de38b02024-06-25 18:25:10 -0700424void VisibilityBasedPreconditioner::RightMultiplyAndAccumulate(
425 const double* x, double* y) const {
Austin Schuh70cc9552019-01-21 19:46:48 -0800426 CHECK(x != nullptr);
427 CHECK(y != nullptr);
428 CHECK(sparse_cholesky_ != nullptr);
429 std::string message;
430 sparse_cholesky_->Solve(x, y, &message);
431}
432
433int VisibilityBasedPreconditioner::num_rows() const { return m_->num_rows(); }
434
435// Classify camera/f_block pairs as in and out of the preconditioner,
436// based on whether the cluster pair that they belong to is in the
437// preconditioner or not.
438bool VisibilityBasedPreconditioner::IsBlockPairInPreconditioner(
439 const int block1, const int block2) const {
440 int cluster1 = cluster_membership_[block1];
441 int cluster2 = cluster_membership_[block2];
442 if (cluster1 > cluster2) {
Austin Schuh3de38b02024-06-25 18:25:10 -0700443 std::swap(cluster1, cluster2);
Austin Schuh70cc9552019-01-21 19:46:48 -0800444 }
Austin Schuh3de38b02024-06-25 18:25:10 -0700445 return (cluster_pairs_.count(std::make_pair(cluster1, cluster2)) > 0);
Austin Schuh70cc9552019-01-21 19:46:48 -0800446}
447
448bool VisibilityBasedPreconditioner::IsBlockPairOffDiagonal(
449 const int block1, const int block2) const {
450 return (cluster_membership_[block1] != cluster_membership_[block2]);
451}
452
453// Convert a graph into a list of edges that includes self edges for
454// each vertex.
455void VisibilityBasedPreconditioner::ForestToClusterPairs(
456 const WeightedGraph<int>& forest,
Austin Schuh3de38b02024-06-25 18:25:10 -0700457 std::unordered_set<std::pair<int, int>, pair_hash>* cluster_pairs) const {
Austin Schuh70cc9552019-01-21 19:46:48 -0800458 CHECK(cluster_pairs != nullptr);
459 cluster_pairs->clear();
460 const std::unordered_set<int>& vertices = forest.vertices();
461 CHECK_EQ(vertices.size(), num_clusters_);
462
463 // Add all the cluster pairs corresponding to the edges in the
464 // forest.
465 for (const int cluster1 : vertices) {
Austin Schuh3de38b02024-06-25 18:25:10 -0700466 cluster_pairs->insert(std::make_pair(cluster1, cluster1));
Austin Schuh70cc9552019-01-21 19:46:48 -0800467 const std::unordered_set<int>& neighbors = forest.Neighbors(cluster1);
468 for (const int cluster2 : neighbors) {
469 if (cluster1 < cluster2) {
Austin Schuh3de38b02024-06-25 18:25:10 -0700470 cluster_pairs->insert(std::make_pair(cluster1, cluster2));
Austin Schuh70cc9552019-01-21 19:46:48 -0800471 }
472 }
473 }
474}
475
476// The visibility set of a cluster is the union of the visibility sets
477// of all its cameras. In other words, the set of points visible to
478// any camera in the cluster.
479void VisibilityBasedPreconditioner::ComputeClusterVisibility(
Austin Schuh3de38b02024-06-25 18:25:10 -0700480 const std::vector<std::set<int>>& visibility,
481 std::vector<std::set<int>>* cluster_visibility) const {
Austin Schuh70cc9552019-01-21 19:46:48 -0800482 CHECK(cluster_visibility != nullptr);
483 cluster_visibility->resize(0);
484 cluster_visibility->resize(num_clusters_);
485 for (int i = 0; i < num_blocks_; ++i) {
486 const int cluster_id = cluster_membership_[i];
487 (*cluster_visibility)[cluster_id].insert(visibility[i].begin(),
488 visibility[i].end());
489 }
490}
491
492// Construct a graph whose vertices are the clusters, and the edge
493// weights are the number of 3D points visible to cameras in both the
494// vertices.
Austin Schuh3de38b02024-06-25 18:25:10 -0700495std::unique_ptr<WeightedGraph<int>>
496VisibilityBasedPreconditioner::CreateClusterGraph(
497 const std::vector<std::set<int>>& cluster_visibility) const {
498 auto cluster_graph = std::make_unique<WeightedGraph<int>>();
Austin Schuh70cc9552019-01-21 19:46:48 -0800499
500 for (int i = 0; i < num_clusters_; ++i) {
501 cluster_graph->AddVertex(i);
502 }
503
504 for (int i = 0; i < num_clusters_; ++i) {
Austin Schuh3de38b02024-06-25 18:25:10 -0700505 const std::set<int>& cluster_i = cluster_visibility[i];
Austin Schuh70cc9552019-01-21 19:46:48 -0800506 for (int j = i + 1; j < num_clusters_; ++j) {
Austin Schuh3de38b02024-06-25 18:25:10 -0700507 std::vector<int> intersection;
508 const std::set<int>& cluster_j = cluster_visibility[j];
509 std::set_intersection(cluster_i.begin(),
510 cluster_i.end(),
511 cluster_j.begin(),
512 cluster_j.end(),
513 std::back_inserter(intersection));
Austin Schuh70cc9552019-01-21 19:46:48 -0800514
515 if (intersection.size() > 0) {
516 // Clusters interact strongly when they share a large number
517 // of 3D points. The degree-2 maximum spanning forest
518 // algorithm, iterates on the edges in decreasing order of
519 // their weight, which is the number of points shared by the
520 // two cameras that it connects.
521 cluster_graph->AddEdge(i, j, intersection.size());
522 }
523 }
524 }
525 return cluster_graph;
526}
527
528// Canonical views clustering returns a std::unordered_map from vertices to
529// cluster ids. Convert this into a flat array for quick lookup. It is
530// possible that some of the vertices may not be associated with any
531// cluster. In that case, randomly assign them to one of the clusters.
532//
533// The cluster ids can be non-contiguous integers. So as we flatten
534// the membership_map, we also map the cluster ids to a contiguous set
535// of integers so that the cluster ids are in [0, num_clusters_).
536void VisibilityBasedPreconditioner::FlattenMembershipMap(
537 const std::unordered_map<int, int>& membership_map,
Austin Schuh3de38b02024-06-25 18:25:10 -0700538 std::vector<int>* membership_vector) const {
Austin Schuh70cc9552019-01-21 19:46:48 -0800539 CHECK(membership_vector != nullptr);
540 membership_vector->resize(0);
541 membership_vector->resize(num_blocks_, -1);
542
543 std::unordered_map<int, int> cluster_id_to_index;
544 // Iterate over the cluster membership map and update the
545 // cluster_membership_ vector assigning arbitrary cluster ids to
546 // the few cameras that have not been clustered.
547 for (const auto& m : membership_map) {
548 const int camera_id = m.first;
549 int cluster_id = m.second;
550
551 // If the view was not clustered, randomly assign it to one of the
552 // clusters. This preserves the mathematical correctness of the
553 // preconditioner. If there are too many views which are not
554 // clustered, it may lead to some quality degradation though.
555 //
556 // TODO(sameeragarwal): Check if a large number of views have not
557 // been clustered and deal with it?
558 if (cluster_id == -1) {
559 cluster_id = camera_id % num_clusters_;
560 }
561
562 const int index = FindWithDefault(
563 cluster_id_to_index, cluster_id, cluster_id_to_index.size());
564
565 if (index == cluster_id_to_index.size()) {
566 cluster_id_to_index[cluster_id] = index;
567 }
568
569 CHECK_LT(index, num_clusters_);
570 membership_vector->at(camera_id) = index;
571 }
572}
573
Austin Schuh3de38b02024-06-25 18:25:10 -0700574} // namespace ceres::internal