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// This include must come before any #ifndef check on Ceres compile options.
33#include "ceres/internal/port.h"
34
35#ifndef CERES_NO_SUITESPARSE
36
37#include "ceres/canonical_views_clustering.h"
38
39#include "ceres/collections_port.h"
40#include "ceres/graph.h"
41#include "gtest/gtest.h"
42
43namespace ceres {
44namespace internal {
45
46const int kVertexIds[] = {0, 1, 2, 3};
47class CanonicalViewsTest : public ::testing::Test {
48 protected:
49  virtual void SetUp() {
50    // The graph structure is as follows.
51    //
52    // Vertex weights:   0      2      2      0
53    //                   V0-----V1-----V2-----V3
54    // Edge weights:        0.8    0.9    0.3
55    const double kVertexWeights[] = {0.0, 2.0, 2.0, -1.0};
56    for (int i = 0; i < 4; ++i) {
57      graph_.AddVertex(i, kVertexWeights[i]);
58    }
59    // Create self edges.
60    // CanonicalViews requires that every view "sees" itself.
61    for (int i = 0; i < 4; ++i) {
62      graph_.AddEdge(i, i, 1.0);
63    }
64
65    // Create three edges.
66    const double kEdgeWeights[] = {0.8, 0.9, 0.3};
67    for (int i = 0; i < 3; ++i) {
68      // The graph interface is directed, so remember to create both
69      // edges.
70      graph_.AddEdge(kVertexIds[i], kVertexIds[i + 1], kEdgeWeights[i]);
71    }
72  }
73
74  void ComputeClustering() {
75    ComputeCanonicalViewsClustering(options_, graph_, &centers_, &membership_);
76  }
77
78  Graph<int> graph_;
79
80  CanonicalViewsClusteringOptions options_;
81  vector<int> centers_;
82  HashMap<int, int> membership_;
83};
84
85TEST_F(CanonicalViewsTest, ComputeCanonicalViewsTest) {
86  options_.min_views = 0;
87  options_.size_penalty_weight = 0.5;
88  options_.similarity_penalty_weight = 0.0;
89  options_.view_score_weight = 0.0;
90  ComputeClustering();
91
92  // 2 canonical views.
93  EXPECT_EQ(centers_.size(), 2);
94  EXPECT_EQ(centers_[0], kVertexIds[1]);
95  EXPECT_EQ(centers_[1], kVertexIds[3]);
96
97  // Check cluster membership.
98  EXPECT_EQ(FindOrDie(membership_, kVertexIds[0]), 0);
99  EXPECT_EQ(FindOrDie(membership_, kVertexIds[1]), 0);
100  EXPECT_EQ(FindOrDie(membership_, kVertexIds[2]), 0);
101  EXPECT_EQ(FindOrDie(membership_, kVertexIds[3]), 1);
102}
103
104// Increases size penalty so the second canonical view won't be
105// chosen.
106TEST_F(CanonicalViewsTest, SizePenaltyTest) {
107  options_.min_views = 0;
108  options_.size_penalty_weight = 2.0;
109  options_.similarity_penalty_weight = 0.0;
110  options_.view_score_weight = 0.0;
111  ComputeClustering();
112
113  // 1 canonical view.
114  EXPECT_EQ(centers_.size(), 1);
115  EXPECT_EQ(centers_[0], kVertexIds[1]);
116}
117
118
119// Increases view score weight so vertex 2 will be chosen.
120TEST_F(CanonicalViewsTest, ViewScoreTest) {
121  options_.min_views = 0;
122  options_.size_penalty_weight = 0.5;
123  options_.similarity_penalty_weight = 0.0;
124  options_.view_score_weight = 1.0;
125  ComputeClustering();
126
127  // 2 canonical views.
128  EXPECT_EQ(centers_.size(), 2);
129  EXPECT_EQ(centers_[0], kVertexIds[1]);
130  EXPECT_EQ(centers_[1], kVertexIds[2]);
131}
132
133// Increases similarity penalty so vertex 2 won't be chosen despite
134// it's view score.
135TEST_F(CanonicalViewsTest, SimilarityPenaltyTest) {
136  options_.min_views = 0;
137  options_.size_penalty_weight = 0.5;
138  options_.similarity_penalty_weight = 3.0;
139  options_.view_score_weight = 1.0;
140  ComputeClustering();
141
142  // 2 canonical views.
143  EXPECT_EQ(centers_.size(), 1);
144  EXPECT_EQ(centers_[0], kVertexIds[1]);
145}
146
147}  // namespace internal
148}  // namespace ceres
149
150#endif  // CERES_NO_SUITESPARSE
151