OpenCV  4.2.0
Open Source Computer Vision
Classify

Goal

In this tutorial you will learn how to

  • How to extract feature from an image
  • How to extract features from images under a given root path
  • How to make a prediction using reference images and target image

Code

/*
* Software License Agreement (BSD License)
*
* Copyright (c) 2009, Willow Garage, Inc.
* All rights reserved.
*
* 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.
* * Neither the name of Willow Garage, Inc. nor the names of its
* contributors may 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
* COPYRIGHT OWNER 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 <iomanip>
using namespace cv;
using namespace std;
using namespace cv::cnn_3dobj;
static void listDir(const char *path, std::vector<String>& files, bool r)
{
DIR *pDir;
struct dirent *ent;
char childpath[512];
pDir = opendir(path);
memset(childpath, 0, sizeof(childpath));
while ((ent = readdir(pDir)) != NULL)
{
if (ent->d_type & DT_DIR)
{
if (strcmp(ent->d_name, ".") == 0 || strcmp(ent->d_name, "..") == 0 || strcmp(ent->d_name, ".DS_Store") == 0)
{
continue;
}
if (r)
{
sprintf(childpath, "%s/%s", path, ent->d_name);
listDir(childpath,files,false);
}
}
else
{
if (strcmp(ent->d_name, ".DS_Store") != 0)
files.push_back(ent->d_name);
}
}
sort(files.begin(),files.end());
};
static int featureWrite(const Mat &features, const String &fname)
{
ofstream ouF;
ouF.open(fname.c_str(), std::ofstream::binary);
if (!ouF)
{
cerr << "failed to open the file : " << fname << endl;
return 0;
}
for (int r = 0; r < features.rows; r++)
{
ouF.write(reinterpret_cast<const char*>(features.ptr(r)), features.cols*features.elemSize());
}
ouF.close();
return 1;
}
int main(int argc, char** argv)
{
const String keys = "{help | | This sample will extract features from reference images and target image for classification. You can add a mean_file if there little variance in data such as human faces, otherwise it is not so useful}"
"{src_dir | ../data/images_all/ | Source direction of the images ready for being used for extract feature as gallery.}"
"{caffemodel | ../../testdata/cv/3d_triplet_iter_30000.caffemodel | caffe model for feature exrtaction.}"
"{network_forIMG | ../../testdata/cv/3d_triplet_testIMG.prototxt | Network definition file used for extracting feature from a single image and making a classification}"
"{mean_file | no | The mean file generated by Caffe from all gallery images, this could be used for mean value substraction from all images. If you want to use the mean file, you can set this as ../data/images_mean/triplet_mean.binaryproto.}"
"{target_img | ../data/images_all/4_78.png | Path of image waiting to be classified.}"
"{feature_blob | feat | Name of layer which will represent as the feature, in this network, ip1 or feat is well.}"
"{num_candidate | 15 | Number of candidates in gallery as the prediction result.}"
"{device | CPU | Device type: CPU or GPU}"
"{dev_id | 0 | Device id}"
"{gallery_out | 0 | Option on output binary features on gallery images}";
/* get parameters from comand line */
cv::CommandLineParser parser(argc, argv, keys);
parser.about("Feature extraction and classification");
if (parser.has("help"))
{
parser.printMessage();
return 0;
}
String src_dir = parser.get<String>("src_dir");
String caffemodel = parser.get<String>("caffemodel");
String network_forIMG = parser.get<String>("network_forIMG");
String mean_file = parser.get<String>("mean_file");
String target_img = parser.get<String>("target_img");
String feature_blob = parser.get<String>("feature_blob");
int num_candidate = parser.get<int>("num_candidate");
String device = parser.get<String>("device");
int gallery_out = parser.get<int>("gallery_out");
/* Initialize a net work with Device */
std::cout << "Using" << descriptor.getDeviceType() << std::endl;
/* Load net with the caffe trained net work parameter and structure */
if (strcmp(mean_file.c_str(), "no") == 0)
descriptor.loadNet(network_forIMG, caffemodel);
else
descriptor.loadNet(network_forIMG, caffemodel, mean_file);
std::vector<String> name_gallery;
/* List the file names under a given path */
listDir(src_dir.c_str(), name_gallery, false);
if (gallery_out)
{
ofstream namelist_out("gallelist.txt");
/* Writing name of the reference images. */
for (unsigned int i = 0; i < name_gallery.size(); i++)
namelist_out << name_gallery.at(i) << endl;
}
for (unsigned int i = 0; i < name_gallery.size(); i++)
{
name_gallery[i] = src_dir + name_gallery[i];
}
std::vector<cv::Mat> img_gallery;
cv::Mat feature_reference;
for (unsigned int i = 0; i < name_gallery.size(); i++)
{
img_gallery.push_back(cv::imread(name_gallery[i]));
}
/* Extract feature from a set of images */
descriptor.extract(img_gallery, feature_reference, feature_blob);
if (gallery_out)
{
std::cout << std::endl << "---------- Features of gallery images ----------" << std::endl;
/* Print features of the reference images. */
for (int i = 0; i < feature_reference.rows; i++)
std::cout << feature_reference.row(i) << endl;
std::cout << std::endl << "---------- Saving features of gallery images into feature.bin ----------" << std::endl;
featureWrite(feature_reference, "feature.bin");
}
else
{
std::cout << std::endl << "---------- Prediction for " << target_img << " ----------" << std::endl;
cv::Mat img = cv::imread(target_img);
std::cout << std::endl << "---------- Features of gallery images ----------" << std::endl;
std::vector<std::pair<String, float> > prediction;
/* Print features of the reference images. */
for (int i = 0; i < feature_reference.rows; i++)
std::cout << feature_reference.row(i) << endl;
cv::Mat feature_test;
descriptor.extract(img, feature_test, feature_blob);
/* Initialize a matcher which using L2 distance. */
std::vector<std::vector<cv::DMatch> > matches;
/* Have a KNN match on the target and reference images. */
matcher.knnMatch(feature_test, feature_reference, matches, num_candidate);
/* Print feature of the target image waiting to be classified. */
std::cout << std::endl << "---------- Features of target image: " << target_img << "----------" << endl << feature_test << std::endl;
/* Print the top N prediction. */
std::cout << std::endl << "---------- Prediction result(Distance - File Name in Gallery) ----------" << std::endl;
for (size_t i = 0; i < matches[0].size(); ++i)
{
std::cout << i << " - " << std::fixed << std::setprecision(2) << name_gallery[matches[0][i].trainIdx] << " - \"" << matches[0][i].distance << "\"" << std::endl;
}
}
return 0;
}

Explanation

Here is the general structure of the program:

  • Initialize a net work with Device.
  • Load net with the caffe trained net work parameter and structure.
    if (strcmp(mean_file.c_str(), "no") == 0)
    descriptor.loadNet(network_forIMG, caffemodel);
    else
    descriptor.loadNet(network_forIMG, caffemodel, mean_file);
  • List the file names under a given path.
    listDir(src_dir.c_str(), name_gallery, false);
    for (unsigned int i = 0; i < name_gallery.size(); i++)
    {
    name_gallery[i] = src_dir + name_gallery[i];
    }
  • Extract feature from a set of images.
    descriptor.extract(img_gallery, feature_reference, feature_blob);
  • Initialize a matcher which using L2 distance.
    std::vector<std::vector<cv::DMatch> > matches;
  • Have a KNN match on the target and reference images.
    matcher.knnMatch(feature_test, feature_reference, matches, num_candidate);
  • Print features of the reference images.
    std::cout << std::endl << "---------- Features of target image: " << target_img << "----------" << endl << feature_test << std::endl;

Results

cv::Mat::rows
int rows
the number of rows and columns or (-1, -1) when the matrix has more than 2 dimensions
Definition: mat.hpp:2086
cv::NORM_L2
Definition: base.hpp:185
cv::String
std::string String
Definition: cvstd.hpp:150
cv::Mat::ptr
uchar * ptr(int i0=0)
Returns a pointer to the specified matrix row.
cv::Mat::elemSize
size_t elemSize() const
Returns the matrix element size in bytes.
cv::BFMatcher
Brute-force descriptor matcher.
Definition: features2d.hpp:1109
cnn_3dobj.hpp
cv::Mat::row
Mat row(int y) const
Creates a matrix header for the specified matrix row.
cv::imread
Mat imread(const String &filename, int flags=IMREAD_COLOR)
Loads an image from a file.
cv::Mat::cols
int cols
Definition: mat.hpp:2086
cv::cnn_3dobj::descriptorExtractor
Caffe based 3D images descriptor. A class to extract features from an image. The so obtained descript...
Definition: cnn_3dobj.hpp:209
features2d.hpp
cv::sort
void sort(InputArray src, OutputArray dst, int flags)
Sorts each row or each column of a matrix.
cv::cnn_3dobj
Definition: cnn_3dobj.hpp:113
cv::Mat
n-dimensional dense array class
Definition: mat.hpp:791
cv::CommandLineParser
Designed for command line parsing.
Definition: utility.hpp:796
cv
"black box" representation of the file storage associated with a file on disk.
Definition: affine.hpp:51