1// Ceres Solver - A fast non-linear least squares minimizer
2// Copyright 2010, 2011, 2012 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: Sameer Agarwal (sameeragarwal@google.com)
30//         David Gallup (dgallup@google.com)
31
32#ifndef CERES_NO_SUITESPARSE
33
34#include "ceres/canonical_views_clustering.h"
35
36#include "ceres/collections_port.h"
37#include "ceres/graph.h"
38#include "gtest/gtest.h"
39
40namespace ceres {
41namespace internal {
42
43const int kVertexIds[] = {0, 1, 2, 3};
44class CanonicalViewsTest : public ::testing::Test {
45 protected:
46  virtual void SetUp() {
47    // The graph structure is as follows.
48    //
49    // Vertex weights:   0      2      2      0
50    //                   V0-----V1-----V2-----V3
51    // Edge weights:        0.8    0.9    0.3
52    const double kVertexWeights[] = {0.0, 2.0, 2.0, -1.0};
53    for (int i = 0; i < 4; ++i) {
54      graph_.AddVertex(i, kVertexWeights[i]);
55    }
56    // Create self edges.
57    // CanonicalViews requires that every view "sees" itself.
58    for (int i = 0; i < 4; ++i) {
59      graph_.AddEdge(i, i, 1.0);
60    }
61
62    // Create three edges.
63    const double kEdgeWeights[] = {0.8, 0.9, 0.3};
64    for (int i = 0; i < 3; ++i) {
65      // The graph interface is directed, so remember to create both
66      // edges.
67      graph_.AddEdge(kVertexIds[i], kVertexIds[i + 1], kEdgeWeights[i]);
68    }
69  }
70
71  void ComputeClustering() {
72    ComputeCanonicalViewsClustering(graph_, options_, &centers_, &membership_);
73  }
74
75  Graph<int> graph_;
76
77  CanonicalViewsClusteringOptions options_;
78  vector<int> centers_;
79  HashMap<int, int> membership_;
80};
81
82TEST_F(CanonicalViewsTest, ComputeCanonicalViewsTest) {
83  options_.min_views = 0;
84  options_.size_penalty_weight = 0.5;
85  options_.similarity_penalty_weight = 0.0;
86  options_.view_score_weight = 0.0;
87  ComputeClustering();
88
89  // 2 canonical views.
90  EXPECT_EQ(centers_.size(), 2);
91  EXPECT_EQ(centers_[0], kVertexIds[1]);
92  EXPECT_EQ(centers_[1], kVertexIds[3]);
93
94  // Check cluster membership.
95  EXPECT_EQ(FindOrDie(membership_, kVertexIds[0]), 0);
96  EXPECT_EQ(FindOrDie(membership_, kVertexIds[1]), 0);
97  EXPECT_EQ(FindOrDie(membership_, kVertexIds[2]), 0);
98  EXPECT_EQ(FindOrDie(membership_, kVertexIds[3]), 1);
99}
100
101// Increases size penalty so the second canonical view won't be
102// chosen.
103TEST_F(CanonicalViewsTest, SizePenaltyTest) {
104  options_.min_views = 0;
105  options_.size_penalty_weight = 2.0;
106  options_.similarity_penalty_weight = 0.0;
107  options_.view_score_weight = 0.0;
108  ComputeClustering();
109
110  // 1 canonical view.
111  EXPECT_EQ(centers_.size(), 1);
112  EXPECT_EQ(centers_[0], kVertexIds[1]);
113}
114
115
116// Increases view score weight so vertex 2 will be chosen.
117TEST_F(CanonicalViewsTest, ViewScoreTest) {
118  options_.min_views = 0;
119  options_.size_penalty_weight = 0.5;
120  options_.similarity_penalty_weight = 0.0;
121  options_.view_score_weight = 1.0;
122  ComputeClustering();
123
124  // 2 canonical views.
125  EXPECT_EQ(centers_.size(), 2);
126  EXPECT_EQ(centers_[0], kVertexIds[1]);
127  EXPECT_EQ(centers_[1], kVertexIds[2]);
128}
129
130// Increases similarity penalty so vertex 2 won't be chosen despite
131// it's view score.
132TEST_F(CanonicalViewsTest, SimilarityPenaltyTest) {
133  options_.min_views = 0;
134  options_.size_penalty_weight = 0.5;
135  options_.similarity_penalty_weight = 3.0;
136  options_.view_score_weight = 1.0;
137  ComputeClustering();
138
139  // 2 canonical views.
140  EXPECT_EQ(centers_.size(), 1);
141  EXPECT_EQ(centers_[0], kVertexIds[1]);
142}
143
144}  // namespace internal
145}  // namespace ceres
146
147#endif  // CERES_NO_SUITESPARSE
148