1/* Ceres Solver - A fast non-linear least squares minimizer
2 * Copyright 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: mierle@gmail.com (Keir Mierle)
30 *
31 * This is a port of curve_fitting.cc to the minimal C API for Ceres.
32 */
33
34#include <math.h>
35#include <stdio.h>
36#include <string.h>  // For NULL
37#include "ceres/c_api.h"
38
39/* Data generated using the following octave code.
40 *
41 *   randn('seed', 23497);
42 *   m = 0.3;
43 *   c = 0.1;
44 *   x=[0:0.075:5];
45 *   y = exp(m * x + c);
46 *   noise = randn(size(x)) * 0.2;
47 *   y_observed = y + noise;
48 *   data = [x', y_observed'];
49 *
50 */
51
52int num_observations = 67;
53double data[] = {
54  0.000000e+00, 1.133898e+00,
55  7.500000e-02, 1.334902e+00,
56  1.500000e-01, 1.213546e+00,
57  2.250000e-01, 1.252016e+00,
58  3.000000e-01, 1.392265e+00,
59  3.750000e-01, 1.314458e+00,
60  4.500000e-01, 1.472541e+00,
61  5.250000e-01, 1.536218e+00,
62  6.000000e-01, 1.355679e+00,
63  6.750000e-01, 1.463566e+00,
64  7.500000e-01, 1.490201e+00,
65  8.250000e-01, 1.658699e+00,
66  9.000000e-01, 1.067574e+00,
67  9.750000e-01, 1.464629e+00,
68  1.050000e+00, 1.402653e+00,
69  1.125000e+00, 1.713141e+00,
70  1.200000e+00, 1.527021e+00,
71  1.275000e+00, 1.702632e+00,
72  1.350000e+00, 1.423899e+00,
73  1.425000e+00, 1.543078e+00,
74  1.500000e+00, 1.664015e+00,
75  1.575000e+00, 1.732484e+00,
76  1.650000e+00, 1.543296e+00,
77  1.725000e+00, 1.959523e+00,
78  1.800000e+00, 1.685132e+00,
79  1.875000e+00, 1.951791e+00,
80  1.950000e+00, 2.095346e+00,
81  2.025000e+00, 2.361460e+00,
82  2.100000e+00, 2.169119e+00,
83  2.175000e+00, 2.061745e+00,
84  2.250000e+00, 2.178641e+00,
85  2.325000e+00, 2.104346e+00,
86  2.400000e+00, 2.584470e+00,
87  2.475000e+00, 1.914158e+00,
88  2.550000e+00, 2.368375e+00,
89  2.625000e+00, 2.686125e+00,
90  2.700000e+00, 2.712395e+00,
91  2.775000e+00, 2.499511e+00,
92  2.850000e+00, 2.558897e+00,
93  2.925000e+00, 2.309154e+00,
94  3.000000e+00, 2.869503e+00,
95  3.075000e+00, 3.116645e+00,
96  3.150000e+00, 3.094907e+00,
97  3.225000e+00, 2.471759e+00,
98  3.300000e+00, 3.017131e+00,
99  3.375000e+00, 3.232381e+00,
100  3.450000e+00, 2.944596e+00,
101  3.525000e+00, 3.385343e+00,
102  3.600000e+00, 3.199826e+00,
103  3.675000e+00, 3.423039e+00,
104  3.750000e+00, 3.621552e+00,
105  3.825000e+00, 3.559255e+00,
106  3.900000e+00, 3.530713e+00,
107  3.975000e+00, 3.561766e+00,
108  4.050000e+00, 3.544574e+00,
109  4.125000e+00, 3.867945e+00,
110  4.200000e+00, 4.049776e+00,
111  4.275000e+00, 3.885601e+00,
112  4.350000e+00, 4.110505e+00,
113  4.425000e+00, 4.345320e+00,
114  4.500000e+00, 4.161241e+00,
115  4.575000e+00, 4.363407e+00,
116  4.650000e+00, 4.161576e+00,
117  4.725000e+00, 4.619728e+00,
118  4.800000e+00, 4.737410e+00,
119  4.875000e+00, 4.727863e+00,
120  4.950000e+00, 4.669206e+00,
121};
122
123/* This is the equivalent of a use-defined CostFunction in the C++ Ceres API.
124 * This is passed as a callback to the Ceres C API, which internally converts
125 * the callback into a CostFunction. */
126int exponential_residual(void* user_data,
127                         double** parameters,
128                         double* residuals,
129                         double** jacobians) {
130  double* measurement = (double*) user_data;
131  double x = measurement[0];
132  double y = measurement[1];
133  double m = parameters[0][0];
134  double c = parameters[1][0];
135
136  residuals[0] = y - exp(m * x + c);
137  if (jacobians == NULL) {
138    return 1;
139  }
140  if (jacobians[0] != NULL) {
141    jacobians[0][0] = - x * exp(m * x + c);  /* dr/dm */
142  }
143  if (jacobians[1] != NULL) {
144    jacobians[1][0] =     - exp(m * x + c);  /* dr/dc */
145  }
146  return 1;
147}
148
149int main(int argc, char** argv) {
150  /* Note: Typically it is better to compact m and c into one block,
151   * but in this case use separate blocks to illustrate the use of
152   * multiple parameter blocks. */
153  double m = 0.0;
154  double c = 0.0;
155
156  double *parameter_pointers[] = { &m, &c };
157  int parameter_sizes[] = { 1, 1 };
158  int i;
159
160  ceres_problem_t* problem;
161
162  /* Ceres has some internal stuff that needs to get initialized. */
163  ceres_init();
164
165  problem = ceres_create_problem();
166
167  /* Add all the residuals. */
168  for (i = 0; i < num_observations; ++i) {
169    ceres_problem_add_residual_block(
170        problem,
171        exponential_residual,  /* Cost function */
172        &data[2 * i],          /* Points to the (x,y) measurement */
173        NULL,                  /* No loss function */
174        NULL,                  /* No loss function user data */
175        1,                     /* Number of residuals */
176        2,                     /* Number of parameter blocks */
177        parameter_sizes,
178        parameter_pointers);
179  }
180
181  ceres_solve(problem);
182  ceres_free_problem(problem);
183
184  printf("Initial m: 0.0, c: 0.0\n");
185  printf("Final m: %g, c: %g\n", m, c);
186  return 0;
187}
188