• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
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: keir@google.com (Keir Mierle)
30 
31 #include "ceres/gradient_checking_cost_function.h"
32 
33 #include <algorithm>
34 #include <cmath>
35 #include <numeric>
36 #include <string>
37 #include <vector>
38 
39 #include "ceres/cost_function.h"
40 #include "ceres/internal/eigen.h"
41 #include "ceres/internal/scoped_ptr.h"
42 #include "ceres/parameter_block.h"
43 #include "ceres/problem.h"
44 #include "ceres/problem_impl.h"
45 #include "ceres/program.h"
46 #include "ceres/residual_block.h"
47 #include "ceres/dynamic_numeric_diff_cost_function.h"
48 #include "ceres/stringprintf.h"
49 #include "ceres/types.h"
50 #include "glog/logging.h"
51 
52 namespace ceres {
53 namespace internal {
54 namespace {
55 
56 // True if x and y have an absolute relative difference less than
57 // relative_precision and false otherwise. Stores the relative and absolute
58 // difference in relative/absolute_error if non-NULL.
IsClose(double x,double y,double relative_precision,double * relative_error,double * absolute_error)59 bool IsClose(double x, double y, double relative_precision,
60              double *relative_error,
61              double *absolute_error) {
62   double local_absolute_error;
63   double local_relative_error;
64   if (!absolute_error) {
65     absolute_error = &local_absolute_error;
66   }
67   if (!relative_error) {
68     relative_error = &local_relative_error;
69   }
70   *absolute_error = fabs(x - y);
71   *relative_error = *absolute_error / max(fabs(x), fabs(y));
72   if (x == 0 || y == 0) {
73     // If x or y is exactly zero, then relative difference doesn't have any
74     // meaning. Take the absolute difference instead.
75     *relative_error = *absolute_error;
76   }
77   return fabs(*relative_error) < fabs(relative_precision);
78 }
79 
80 class GradientCheckingCostFunction : public CostFunction {
81  public:
GradientCheckingCostFunction(const CostFunction * function,double relative_step_size,double relative_precision,const string & extra_info)82   GradientCheckingCostFunction(const CostFunction* function,
83                                double relative_step_size,
84                                double relative_precision,
85                                const string& extra_info)
86       : function_(function),
87         relative_precision_(relative_precision),
88         extra_info_(extra_info) {
89     DynamicNumericDiffCostFunction<CostFunction, CENTRAL>*
90         finite_diff_cost_function =
91         new DynamicNumericDiffCostFunction<CostFunction, CENTRAL>(
92             function,
93             DO_NOT_TAKE_OWNERSHIP,
94             relative_step_size);
95 
96     const vector<int32>& parameter_block_sizes =
97         function->parameter_block_sizes();
98     for (int i = 0; i < parameter_block_sizes.size(); ++i) {
99       finite_diff_cost_function->AddParameterBlock(parameter_block_sizes[i]);
100     }
101     *mutable_parameter_block_sizes() = parameter_block_sizes;
102     set_num_residuals(function->num_residuals());
103     finite_diff_cost_function->SetNumResiduals(num_residuals());
104     finite_diff_cost_function_.reset(finite_diff_cost_function);
105   }
106 
~GradientCheckingCostFunction()107   virtual ~GradientCheckingCostFunction() { }
108 
Evaluate(double const * const * parameters,double * residuals,double ** jacobians) const109   virtual bool Evaluate(double const* const* parameters,
110                         double* residuals,
111                         double** jacobians) const {
112     if (!jacobians) {
113       // Nothing to check in this case; just forward.
114       return function_->Evaluate(parameters, residuals, NULL);
115     }
116 
117     int num_residuals = function_->num_residuals();
118 
119     // Make space for the jacobians of the two methods.
120     const vector<int32>& block_sizes = function_->parameter_block_sizes();
121     vector<Matrix> term_jacobians(block_sizes.size());
122     vector<Matrix> finite_difference_jacobians(block_sizes.size());
123     vector<double*> term_jacobian_pointers(block_sizes.size());
124     vector<double*> finite_difference_jacobian_pointers(block_sizes.size());
125     for (int i = 0; i < block_sizes.size(); i++) {
126       term_jacobians[i].resize(num_residuals, block_sizes[i]);
127       term_jacobian_pointers[i] = term_jacobians[i].data();
128       finite_difference_jacobians[i].resize(num_residuals, block_sizes[i]);
129       finite_difference_jacobian_pointers[i] =
130           finite_difference_jacobians[i].data();
131     }
132 
133     // Evaluate the derivative using the user supplied code.
134     if (!function_->Evaluate(parameters,
135                              residuals,
136                              &term_jacobian_pointers[0])) {
137       LOG(WARNING) << "Function evaluation failed.";
138       return false;
139     }
140 
141     // Evaluate the derivative using numeric derivatives.
142     finite_diff_cost_function_->Evaluate(
143         parameters,
144         residuals,
145         &finite_difference_jacobian_pointers[0]);
146 
147     // See if any elements have relative error larger than the threshold.
148     int num_bad_jacobian_components = 0;
149     double worst_relative_error = 0;
150 
151     // Accumulate the error message for all the jacobians, since it won't get
152     // output if there are no bad jacobian components.
153     string m;
154     for (int k = 0; k < block_sizes.size(); k++) {
155       // Copy the original jacobian blocks into the jacobians array.
156       if (jacobians[k] != NULL) {
157         MatrixRef(jacobians[k],
158                   term_jacobians[k].rows(),
159                   term_jacobians[k].cols()) = term_jacobians[k];
160       }
161 
162       StringAppendF(&m,
163                     "========== "
164                     "Jacobian for " "block %d: (%ld by %ld)) "
165                     "==========\n",
166                     k,
167                     static_cast<long>(term_jacobians[k].rows()),
168                     static_cast<long>(term_jacobians[k].cols()));
169       // The funny spacing creates appropriately aligned column headers.
170       m += " block  row  col        user dx/dy    num diff dx/dy         "
171            "abs error    relative error         parameter          residual\n";
172 
173       for (int i = 0; i < term_jacobians[k].rows(); i++) {
174         for (int j = 0; j < term_jacobians[k].cols(); j++) {
175           double term_jacobian = term_jacobians[k](i, j);
176           double finite_jacobian = finite_difference_jacobians[k](i, j);
177           double relative_error, absolute_error;
178           bool bad_jacobian_entry =
179               !IsClose(term_jacobian,
180                        finite_jacobian,
181                        relative_precision_,
182                        &relative_error,
183                        &absolute_error);
184           worst_relative_error = std::max(worst_relative_error,
185                                           relative_error);
186 
187           StringAppendF(&m, "%6d %4d %4d %17g %17g %17g %17g %17g %17g",
188                         k, i, j,
189                         term_jacobian, finite_jacobian,
190                         absolute_error, relative_error,
191                         parameters[k][j],
192                         residuals[i]);
193 
194           if (bad_jacobian_entry) {
195             num_bad_jacobian_components++;
196             StringAppendF(
197                 &m, " ------ (%d,%d,%d) Relative error worse than %g",
198                 k, i, j, relative_precision_);
199           }
200           m += "\n";
201         }
202       }
203     }
204 
205     // Since there were some bad errors, dump comprehensive debug info.
206     if (num_bad_jacobian_components) {
207       string header = StringPrintf("Detected %d bad jacobian component(s). "
208                                    "Worst relative error was %g.\n",
209                                    num_bad_jacobian_components,
210                                    worst_relative_error);
211       if (!extra_info_.empty()) {
212         header += "Extra info for this residual: " + extra_info_ + "\n";
213       }
214       LOG(WARNING) << "\n" << header << m;
215     }
216     return true;
217   }
218 
219  private:
220   const CostFunction* function_;
221   internal::scoped_ptr<CostFunction> finite_diff_cost_function_;
222   double relative_precision_;
223   string extra_info_;
224 };
225 
226 }  // namespace
227 
CreateGradientCheckingCostFunction(const CostFunction * cost_function,double relative_step_size,double relative_precision,const string & extra_info)228 CostFunction *CreateGradientCheckingCostFunction(
229     const CostFunction *cost_function,
230     double relative_step_size,
231     double relative_precision,
232     const string& extra_info) {
233   return new GradientCheckingCostFunction(cost_function,
234                                           relative_step_size,
235                                           relative_precision,
236                                           extra_info);
237 }
238 
CreateGradientCheckingProblemImpl(ProblemImpl * problem_impl,double relative_step_size,double relative_precision)239 ProblemImpl* CreateGradientCheckingProblemImpl(ProblemImpl* problem_impl,
240                                                double relative_step_size,
241                                                double relative_precision) {
242   // We create new CostFunctions by wrapping the original CostFunction
243   // in a gradient checking CostFunction. So its okay for the
244   // ProblemImpl to take ownership of it and destroy it. The
245   // LossFunctions and LocalParameterizations are reused and since
246   // they are owned by problem_impl, gradient_checking_problem_impl
247   // should not take ownership of it.
248   Problem::Options gradient_checking_problem_options;
249   gradient_checking_problem_options.cost_function_ownership = TAKE_OWNERSHIP;
250   gradient_checking_problem_options.loss_function_ownership =
251       DO_NOT_TAKE_OWNERSHIP;
252   gradient_checking_problem_options.local_parameterization_ownership =
253       DO_NOT_TAKE_OWNERSHIP;
254 
255   ProblemImpl* gradient_checking_problem_impl = new ProblemImpl(
256       gradient_checking_problem_options);
257 
258   Program* program = problem_impl->mutable_program();
259 
260   // For every ParameterBlock in problem_impl, create a new parameter
261   // block with the same local parameterization and constancy.
262   const vector<ParameterBlock*>& parameter_blocks = program->parameter_blocks();
263   for (int i = 0; i < parameter_blocks.size(); ++i) {
264     ParameterBlock* parameter_block = parameter_blocks[i];
265     gradient_checking_problem_impl->AddParameterBlock(
266         parameter_block->mutable_user_state(),
267         parameter_block->Size(),
268         parameter_block->mutable_local_parameterization());
269 
270     if (parameter_block->IsConstant()) {
271       gradient_checking_problem_impl->SetParameterBlockConstant(
272           parameter_block->mutable_user_state());
273     }
274   }
275 
276   // For every ResidualBlock in problem_impl, create a new
277   // ResidualBlock by wrapping its CostFunction inside a
278   // GradientCheckingCostFunction.
279   const vector<ResidualBlock*>& residual_blocks = program->residual_blocks();
280   for (int i = 0; i < residual_blocks.size(); ++i) {
281     ResidualBlock* residual_block = residual_blocks[i];
282 
283     // Build a human readable string which identifies the
284     // ResidualBlock. This is used by the GradientCheckingCostFunction
285     // when logging debugging information.
286     string extra_info = StringPrintf(
287         "Residual block id %d; depends on parameters [", i);
288     vector<double*> parameter_blocks;
289     for (int j = 0; j < residual_block->NumParameterBlocks(); ++j) {
290       ParameterBlock* parameter_block = residual_block->parameter_blocks()[j];
291       parameter_blocks.push_back(parameter_block->mutable_user_state());
292       StringAppendF(&extra_info, "%p", parameter_block->mutable_user_state());
293       extra_info += (j < residual_block->NumParameterBlocks() - 1) ? ", " : "]";
294     }
295 
296     // Wrap the original CostFunction in a GradientCheckingCostFunction.
297     CostFunction* gradient_checking_cost_function =
298         CreateGradientCheckingCostFunction(residual_block->cost_function(),
299                                            relative_step_size,
300                                            relative_precision,
301                                            extra_info);
302 
303     // The const_cast is necessary because
304     // ProblemImpl::AddResidualBlock can potentially take ownership of
305     // the LossFunction, but in this case we are guaranteed that this
306     // will not be the case, so this const_cast is harmless.
307     gradient_checking_problem_impl->AddResidualBlock(
308         gradient_checking_cost_function,
309         const_cast<LossFunction*>(residual_block->loss_function()),
310         parameter_blocks);
311   }
312 
313   return gradient_checking_problem_impl;
314 }
315 
316 
317 }  // namespace internal
318 }  // namespace ceres
319