potok/opencv
1
0
Fork 0
mirror of https://github.com/opencv/opencv.git synced 2026-07-30 07:43:03 +04:00
Code Packages Releases Wiki Activity

Merge pull request #15959 from mshabunin:refactor-ml-tests

Browse Source 
ml: refactored tests

* use parametrized tests where appropriate
* use stable theRNG in most tests
* use modern style with EXPECT_/ASSERT_ checks
This commit is contained in:
Maksim Shabunin
2019-11-25 20:03:16 +00:00
committed by Alexander Alekhin
parent 9e906d9e21
commit 5ff1fababc
16 changed files with 1418 additions and 2857 deletions
Download Patch File Download Diff File Expand all files Collapse all files
modules/ml/test/test_lr.cpp
+23 -169
View File
@@ -1,9 +1,6 @@
///////////////////////////////////////////////////////////////////////////////////////
// IMPORTANT: READ BEFORE DOWNLOADING, COPYING, INSTALLING OR USING.
// By downloading, copying, installing or using the software you agree to this license.
// If you do not agree to this license, do not download, install,
// copy or use the software.
// This file is part of OpenCV project.
// It is subject to the license terms in the LICENSE file found in the top-level directory
// of this distribution and at http://opencv.org/license.html.
// This is a implementation of the Logistic Regression algorithm in C++ in OpenCV.
@@ -11,92 +8,16 @@
// Rahul Kavi rahulkavi[at]live[at]com
//
// contains a subset of data from the popular Iris Dataset (taken from "http://archive.ics.uci.edu/ml/datasets/Iris")
// # You are free to use, change, or redistribute the code in any way you wish for
// # non-commercial purposes, but please maintain the name of the original author.
// # This code comes with no warranty of any kind.
// #
// # You are free to use, change, or redistribute the code in any way you wish for
// # non-commercial purposes, but please maintain the name of the original author.
// # This code comes with no warranty of any kind.
// # Logistic Regression ALGORITHM
// License Agreement
// For Open Source Computer Vision Library
// Copyright (C) 2000-2008, Intel Corporation, all rights reserved.
// Copyright (C) 2008-2011, Willow Garage Inc., all rights reserved.
// Third party copyrights are property of their respective owners.
// Redistribution and use in source and binary forms, with or without modification,
// are permitted provided that the following conditions are met:
// * Redistributions of source code must retain the above copyright notice,
// this list of conditions and the following disclaimer.
// * Redistributions in binary form must reproduce the above copyright notice,
// this list of conditions and the following disclaimer in the documentation
// and/or other materials provided with the distribution.
// * The name of the copyright holders may not be used to endorse or promote products
// derived from this software without specific prior written permission.
// This software is provided by the copyright holders and contributors "as is" and
// any express or implied warranties, including, but not limited to, the implied
// warranties of merchantability and fitness for a particular purpose are disclaimed.
// In no event shall the Intel Corporation or contributors be liable for any direct,
// indirect, incidental, special, exemplary, or consequential damages
// (including, but not limited to, procurement of substitute goods or services;
// loss of use, data, or profits; or business interruption) however caused
// and on any theory of liability, whether in contract, strict liability,
// or tort (including negligence or otherwise) arising in any way out of
// the use of this software, even if advised of the possibility of such damage.
#include "test_precomp.hpp"
namespace opencv_test { namespace {
bool calculateError( const Mat& _p_labels, const Mat& _o_labels, float& error)
TEST(ML_LR, accuracy)
{
CV_TRACE_FUNCTION();
error = 0.0f;
float accuracy = 0.0f;
Mat _p_labels_temp;
Mat _o_labels_temp;
_p_labels.convertTo(_p_labels_temp, CV_32S);
_o_labels.convertTo(_o_labels_temp, CV_32S);
CV_Assert(_p_labels_temp.total() == _o_labels_temp.total());
CV_Assert(_p_labels_temp.rows == _o_labels_temp.rows);
accuracy = (float)countNonZero(_p_labels_temp == _o_labels_temp)/_p_labels_temp.rows;
error = 1 - accuracy;
return true;
}
//--------------------------------------------------------------------------------------------
class CV_LRTest : public cvtest::BaseTest
{
public:
CV_LRTest() {}
protected:
virtual void run( int start_from );
};
void CV_LRTest::run( int /*start_from*/ )
{
CV_TRACE_FUNCTION();
// initialize variables from the popular Iris Dataset
string dataFileName = ts->get_data_path() + "iris.data";
std::string dataFileName = findDataFile("iris.data");
Ptr<TrainData> tdata = TrainData::loadFromCSV(dataFileName, 0);
ASSERT_FALSE(tdata.empty()) << "Could not find test data file : " << dataFileName;
ASSERT_FALSE(tdata.empty());
// run LR classifier train classifier
Ptr<LogisticRegression> p = LogisticRegression::create();
p->setLearningRate(1.0);
p->setIterations(10001);
@@ -105,121 +26,54 @@ void CV_LRTest::run( int /*start_from*/ )
p->setMiniBatchSize(10);
p->train(tdata);
// predict using the same data
Mat responses;
p->predict(tdata->getSamples(), responses);
// calculate error
int test_code = cvtest::TS::OK;
float error = 0.0f;
if(!calculateError(responses, tdata->getResponses(), error))
{
ts->printf(cvtest::TS::LOG, "Bad prediction labels\n" );
test_code = cvtest::TS::FAIL_INVALID_OUTPUT;
}
else if(error > 0.05f)
{
ts->printf(cvtest::TS::LOG, "Bad accuracy of (%f)\n", error);
test_code = cvtest::TS::FAIL_BAD_ACCURACY;
}
{
FileStorage s("debug.xml", FileStorage::WRITE);
s << "original" << tdata->getResponses();
s << "predicted1" << responses;
s << "learnt" << p->get_learnt_thetas();
s << "error" << error;
s.release();
}
ts->set_failed_test_info(test_code);
float error = 1000;
EXPECT_TRUE(calculateError(responses, tdata->getResponses(), error));
EXPECT_LE(error, 0.05f);
}
//--------------------------------------------------------------------------------------------
class CV_LRTest_SaveLoad : public cvtest::BaseTest
//==================================================================================================
TEST(ML_LR, save_load)
{
public:
CV_LRTest_SaveLoad(){}
protected:
virtual void run(int start_from);
};
void CV_LRTest_SaveLoad::run( int /*start_from*/ )
{
CV_TRACE_FUNCTION();
int code = cvtest::TS::OK;
// initialize variables from the popular Iris Dataset
string dataFileName = ts->get_data_path() + "iris.data";
string dataFileName = findDataFile("iris.data");
Ptr<TrainData> tdata = TrainData::loadFromCSV(dataFileName, 0);
ASSERT_FALSE(tdata.empty()) << "Could not find test data file : " << dataFileName;
ASSERT_FALSE(tdata.empty());
Mat responses1, responses2;
Mat learnt_mat1, learnt_mat2;
// train and save the classifier
String filename = tempfile(".xml");
try
{
// run LR classifier train classifier
Ptr<LogisticRegression> lr1 = LogisticRegression::create();
lr1->setLearningRate(1.0);
lr1->setIterations(10001);
lr1->setRegularization(LogisticRegression::REG_L2);
lr1->setTrainMethod(LogisticRegression::BATCH);
lr1->setMiniBatchSize(10);
lr1->train(tdata);
lr1->predict(tdata->getSamples(), responses1);
ASSERT_NO_THROW(lr1->train(tdata));
ASSERT_NO_THROW(lr1->predict(tdata->getSamples(), responses1));
ASSERT_NO_THROW(lr1->save(filename));
learnt_mat1 = lr1->get_learnt_thetas();
lr1->save(filename);
}
catch(...)
{
ts->printf(cvtest::TS::LOG, "Crash in write method.\n" );
ts->set_failed_test_info(cvtest::TS::FAIL_EXCEPTION);
}
// and load to another
try
{
Ptr<LogisticRegression> lr2 = Algorithm::load<LogisticRegression>(filename);
lr2->predict(tdata->getSamples(), responses2);
Ptr<LogisticRegression> lr2;
ASSERT_NO_THROW(lr2 = Algorithm::load<LogisticRegression>(filename));
ASSERT_NO_THROW(lr2->predict(tdata->getSamples(), responses2));
learnt_mat2 = lr2->get_learnt_thetas();
}
catch(...)
{
ts->printf(cvtest::TS::LOG, "Crash in write method.\n" );
ts->set_failed_test_info(cvtest::TS::FAIL_EXCEPTION);
}
// compare difference in prediction outputs and stored inputs
EXPECT_MAT_NEAR(responses1, responses2, 0.f);
CV_Assert(responses1.rows == responses2.rows);
// compare difference in learnt matrices before and after loading from disk
Mat comp_learnt_mats;
comp_learnt_mats = (learnt_mat1 == learnt_mat2);
comp_learnt_mats = comp_learnt_mats.reshape(1, comp_learnt_mats.rows*comp_learnt_mats.cols);
comp_learnt_mats.convertTo(comp_learnt_mats, CV_32S);
comp_learnt_mats = comp_learnt_mats/255;
// compare difference in prediction outputs and stored inputs
// check if there is any difference between computed learnt mat and retrieved mat
float errorCount = 0.0;
errorCount += 1 - (float)countNonZero(responses1 == responses2)/responses1.rows;
errorCount += 1 - (float)sum(comp_learnt_mats)[0]/comp_learnt_mats.rows;
if(errorCount>0)
{
ts->printf( cvtest::TS::LOG, "Different prediction results before writing and after reading (errorCount=%d).\n", errorCount );
code = cvtest::TS::FAIL_BAD_ACCURACY;
}
EXPECT_EQ(comp_learnt_mats.rows, sum(comp_learnt_mats)[0]);
remove( filename.c_str() );
ts->set_failed_test_info( code );
}
TEST(ML_LR, accuracy) { CV_LRTest test; test.safe_run(); }
TEST(ML_LR, save_load) { CV_LRTest_SaveLoad test; test.safe_run(); }
}} // namespace