• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1 // Ceres Solver - A fast non-linear least squares minimizer
2 // Copyright 2010, 2011, 2012, 2013 Google Inc. All rights reserved.
3 // http://code.google.com/p/ceres-solver/
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: keir@google.com (Keir Mierle)
30 //
31 // TODO(keir): Implement a generic "compare sparse matrix implementations" test
32 // suite that can compare all the implementations. Then this file would shrink
33 // in size.
34 
35 #include "ceres/dense_sparse_matrix.h"
36 
37 #include "ceres/casts.h"
38 #include "ceres/linear_least_squares_problems.h"
39 #include "ceres/triplet_sparse_matrix.h"
40 #include "ceres/internal/eigen.h"
41 #include "ceres/internal/scoped_ptr.h"
42 #include "glog/logging.h"
43 #include "gtest/gtest.h"
44 
45 namespace ceres {
46 namespace internal {
47 
CompareMatrices(const SparseMatrix * a,const SparseMatrix * b)48 void CompareMatrices(const SparseMatrix* a, const SparseMatrix* b) {
49   EXPECT_EQ(a->num_rows(), b->num_rows());
50   EXPECT_EQ(a->num_cols(), b->num_cols());
51 
52   int num_rows = a->num_rows();
53   int num_cols = a->num_cols();
54 
55   for (int i = 0; i < num_cols; ++i) {
56     Vector x = Vector::Zero(num_cols);
57     x(i) = 1.0;
58 
59     Vector y_a = Vector::Zero(num_rows);
60     Vector y_b = Vector::Zero(num_rows);
61 
62     a->RightMultiply(x.data(), y_a.data());
63     b->RightMultiply(x.data(), y_b.data());
64 
65     EXPECT_EQ((y_a - y_b).norm(), 0);
66   }
67 }
68 
69 class DenseSparseMatrixTest : public ::testing::Test {
70  protected :
SetUp()71   virtual void SetUp() {
72     scoped_ptr<LinearLeastSquaresProblem> problem(
73         CreateLinearLeastSquaresProblemFromId(1));
74 
75     CHECK_NOTNULL(problem.get());
76 
77     tsm.reset(down_cast<TripletSparseMatrix*>(problem->A.release()));
78     dsm.reset(new DenseSparseMatrix(*tsm));
79 
80     num_rows = tsm->num_rows();
81     num_cols = tsm->num_cols();
82   }
83 
84   int num_rows;
85   int num_cols;
86 
87   scoped_ptr<TripletSparseMatrix> tsm;
88   scoped_ptr<DenseSparseMatrix> dsm;
89 };
90 
TEST_F(DenseSparseMatrixTest,RightMultiply)91 TEST_F(DenseSparseMatrixTest, RightMultiply) {
92   CompareMatrices(tsm.get(), dsm.get());
93 
94   // Try with a not entirely zero vector to verify column interactions, which
95   // could be masked by a subtle bug when using the elementary vectors.
96   Vector a(num_cols);
97   for (int i = 0; i < num_cols; i++) {
98     a(i) = i;
99   }
100   Vector b1 = Vector::Zero(num_rows);
101   Vector b2 = Vector::Zero(num_rows);
102 
103   tsm->RightMultiply(a.data(), b1.data());
104   dsm->RightMultiply(a.data(), b2.data());
105 
106   EXPECT_EQ((b1 - b2).norm(), 0);
107 }
108 
TEST_F(DenseSparseMatrixTest,LeftMultiply)109 TEST_F(DenseSparseMatrixTest, LeftMultiply) {
110   for (int i = 0; i < num_rows; ++i) {
111     Vector a = Vector::Zero(num_rows);
112     a(i) = 1.0;
113 
114     Vector b1 = Vector::Zero(num_cols);
115     Vector b2 = Vector::Zero(num_cols);
116 
117     tsm->LeftMultiply(a.data(), b1.data());
118     dsm->LeftMultiply(a.data(), b2.data());
119 
120     EXPECT_EQ((b1 - b2).norm(), 0);
121   }
122 
123   // Try with a not entirely zero vector to verify column interactions, which
124   // could be masked by a subtle bug when using the elementary vectors.
125   Vector a(num_rows);
126   for (int i = 0; i < num_rows; i++) {
127     a(i) = i;
128   }
129   Vector b1 = Vector::Zero(num_cols);
130   Vector b2 = Vector::Zero(num_cols);
131 
132   tsm->LeftMultiply(a.data(), b1.data());
133   dsm->LeftMultiply(a.data(), b2.data());
134 
135   EXPECT_EQ((b1 - b2).norm(), 0);
136 }
137 
TEST_F(DenseSparseMatrixTest,ColumnNorm)138 TEST_F(DenseSparseMatrixTest, ColumnNorm) {
139   Vector b1 = Vector::Zero(num_cols);
140   Vector b2 = Vector::Zero(num_cols);
141 
142   tsm->SquaredColumnNorm(b1.data());
143   dsm->SquaredColumnNorm(b2.data());
144 
145   EXPECT_EQ((b1 - b2).norm(), 0);
146 }
147 
TEST_F(DenseSparseMatrixTest,Scale)148 TEST_F(DenseSparseMatrixTest, Scale) {
149   Vector scale(num_cols);
150   for (int i = 0; i < num_cols; ++i) {
151     scale(i) = i + 1;
152   }
153   tsm->ScaleColumns(scale.data());
154   dsm->ScaleColumns(scale.data());
155   CompareMatrices(tsm.get(), dsm.get());
156 }
157 
TEST_F(DenseSparseMatrixTest,ToDenseMatrix)158 TEST_F(DenseSparseMatrixTest, ToDenseMatrix) {
159   Matrix tsm_dense;
160   Matrix dsm_dense;
161 
162   tsm->ToDenseMatrix(&tsm_dense);
163   dsm->ToDenseMatrix(&dsm_dense);
164 
165   EXPECT_EQ((tsm_dense - dsm_dense).norm(), 0.0);
166 }
167 
168 }  // namespace internal
169 }  // namespace ceres
170