opencv/modules/gpu/test/test_meanshift.cpp

/*M///////////////////////////////////////////////////////////////////////////////////////
//
//  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.
//
//
//                        Intel License Agreement
//                For Open Source Computer Vision Library
//
// Copyright (C) 2000, Intel Corporation, 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:
//
//   * Redistribution's of source code must retain the above copyright notice,
//     this list of conditions and the following disclaimer.
//
//   * Redistribution's 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 Intel Corporation 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.
//
//M*/

#include "test_precomp.hpp"
#include <iostream>
#include <string>

using namespace cv;
using namespace cv::gpu;


struct CV_GpuMeanShiftTest : public cvtest::BaseTest
{
    CV_GpuMeanShiftTest() {}

    void run(int)
    {
        bool cc12_ok = TargetArchs::builtWith(FEATURE_SET_COMPUTE_12) && DeviceInfo().supports(FEATURE_SET_COMPUTE_12);
        if (!cc12_ok)
        {
            ts->printf(cvtest::TS::CONSOLE, "\nCompute capability 1.2 is required");
            ts->set_failed_test_info(cvtest::TS::FAIL_GENERIC);
            return;
        }

        int spatialRad = 30;
        int colorRad = 30;

        cv::Mat img = cv::imread(std::string(ts->get_data_path()) + "meanshift/cones.png");
        cv::Mat img_template;       
        
        if (cv::gpu::TargetArchs::builtWith(cv::gpu::FEATURE_SET_COMPUTE_20) &&
            cv::gpu::DeviceInfo().supports(cv::gpu::FEATURE_SET_COMPUTE_20))
            img_template = cv::imread(std::string(ts->get_data_path()) + "meanshift/con_result.png");
        else
            img_template = cv::imread(std::string(ts->get_data_path()) + "meanshift/con_result_CC1X.png");

        if (img.empty() || img_template.empty())
        {
            ts->set_failed_test_info(cvtest::TS::FAIL_MISSING_TEST_DATA);
            return;
        }

        cv::Mat rgba;
        cvtColor(img, rgba, CV_BGR2BGRA);


        cv::gpu::GpuMat res;
        cv::gpu::meanShiftFiltering( cv::gpu::GpuMat(rgba), res, spatialRad, colorRad );

        if (res.type() != CV_8UC4)
        {
            ts->set_failed_test_info(cvtest::TS::FAIL_INVALID_OUTPUT);
            return;
        }

        cv::Mat result;
        res.download(result);

        uchar maxDiff = 0;
        for (int j = 0; j < result.rows; ++j)
        {
            const uchar* res_line = result.ptr<uchar>(j);
            const uchar* ref_line = img_template.ptr<uchar>(j);

            for (int i = 0; i < result.cols; ++i)
            {
                for (int k = 0; k < 3; ++k)
                {
                    const uchar& ch1 = res_line[result.channels()*i + k];
                    const uchar& ch2 = ref_line[img_template.channels()*i + k];
                    uchar diff = static_cast<uchar>(abs(ch1 - ch2));
                    if (maxDiff < diff)
                        maxDiff = diff;
                }
            }
        }
        if (maxDiff > 0)
        {
            ts->printf(cvtest::TS::LOG, "\nMeanShift maxDiff = %d\n", maxDiff);
            ts->set_failed_test_info(cvtest::TS::FAIL_GENERIC);
            return;
        }

        ts->set_failed_test_info(cvtest::TS::OK);
    }

};

TEST(meanShift, accuracy) { CV_GpuMeanShiftTest test; test.safe_run(); }

struct CV_GpuMeanShiftProcTest : public cvtest::BaseTest
{
    CV_GpuMeanShiftProcTest() {}

    void run(int)
    {
        bool cc12_ok = TargetArchs::builtWith(FEATURE_SET_COMPUTE_12) && DeviceInfo().supports(FEATURE_SET_COMPUTE_12);
        if (!cc12_ok)
        {
            ts->printf(cvtest::TS::CONSOLE, "\nCompute capability 1.2 is required");
            ts->set_failed_test_info(cvtest::TS::FAIL_GENERIC);
            return;
        }

        int spatialRad = 30;
        int colorRad = 30;

        cv::Mat img = cv::imread(std::string(ts->get_data_path()) + "meanshift/cones.png");

        if (img.empty())
        {
            ts->set_failed_test_info(cvtest::TS::FAIL_MISSING_TEST_DATA);
            return;
        }

        cv::Mat rgba;
        cvtColor(img, rgba, CV_BGR2BGRA);

        cv::gpu::GpuMat h_rmap_filtered;
        cv::gpu::meanShiftFiltering( cv::gpu::GpuMat(rgba), h_rmap_filtered, spatialRad, colorRad );

        cv::gpu::GpuMat d_rmap;
        cv::gpu::GpuMat d_spmap;
        cv::gpu::meanShiftProc( cv::gpu::GpuMat(rgba), d_rmap, d_spmap, spatialRad, colorRad );

        if (d_rmap.type() != CV_8UC4)
        {
            ts->set_failed_test_info(cvtest::TS::FAIL_INVALID_OUTPUT);
            return;
        }

        cv::Mat rmap_filtered;
        h_rmap_filtered.download(rmap_filtered);

        cv::Mat rmap;
        d_rmap.download(rmap);

        uchar maxDiff = 0;
        for (int j = 0; j < rmap_filtered.rows; ++j)
        {
            const uchar* res_line = rmap_filtered.ptr<uchar>(j);
            const uchar* ref_line = rmap.ptr<uchar>(j);

            for (int i = 0; i < rmap_filtered.cols; ++i)
            {
                for (int k = 0; k < 3; ++k)
                {
                    const uchar& ch1 = res_line[rmap_filtered.channels()*i + k];
                    const uchar& ch2 = ref_line[rmap.channels()*i + k];
                    uchar diff = static_cast<uchar>(abs(ch1 - ch2));
                    if (maxDiff < diff)
                        maxDiff = diff;
                }
            }
        }
        if (maxDiff > 0)
        {
            ts->printf(cvtest::TS::LOG, "\nMeanShiftProc maxDiff = %d\n", maxDiff);
            ts->set_failed_test_info(cvtest::TS::FAIL_GENERIC);
            return;
        }

        cv::Mat spmap;
        d_spmap.download(spmap);

        cv::Mat spmap_template;
        cv::FileStorage fs;

        if (cv::gpu::TargetArchs::builtWith(cv::gpu::FEATURE_SET_COMPUTE_20) &&
            cv::gpu::DeviceInfo().supports(cv::gpu::FEATURE_SET_COMPUTE_20))
            fs.open(std::string(ts->get_data_path()) + "meanshift/spmap.yaml", cv::FileStorage::READ);
        else
            fs.open(std::string(ts->get_data_path()) + "meanshift/spmap_CC1X.yaml", cv::FileStorage::READ);
        fs["spmap"] >> spmap_template;

        for (int y = 0; y < spmap.rows; ++y) {
            for (int x = 0; x < spmap.cols; ++x) {
                cv::Point_<short> expected = spmap_template.at<cv::Point_<short> >(y, x);
                cv::Point_<short> actual = spmap.at<cv::Point_<short> >(y, x);
                int diff = (expected - actual).dot(expected - actual);
                if (actual != expected) {
                    ts->printf(cvtest::TS::LOG, "\nMeanShiftProc SpMap is bad, diff=%d\n", diff);
                    ts->set_failed_test_info(cvtest::TS::FAIL_GENERIC);
                    return;
                }
            }
        }

        ts->set_failed_test_info(cvtest::TS::OK);
    }

};

TEST(meanShiftProc, accuracy) { CV_GpuMeanShiftProcTest test; test.safe_run(); }