I'm having problems getting PCA and Eigenfaces working using the latest C++ syntax with the Mat and PCA classes. The older C syntax took an array of IplImage* as a parameter to perform its processing and the current API only takes a Mat that is formatted by Column or Row. I took the Row approach using the reshape function to fit my image's matrix to fit in a single row. I eventually want to take this data and then use the SVM algorithm to perform detection, but when I do that all my data is just a stream of 0s. Can someone please help me out? What am I doing wrong? Thanks!
I saw this question and it's somewhat related, but I'm not sure what the solution is.
This is basically what I have:
vector<Mat> images; //This variable will be loaded with a set of images to perform PCA on.
Mat values(images.size(), 1, CV_32SC1); //Values are the corresponding values to each of my images.
int nEigens = images.size() - 1; //Number of Eigen Vectors.
//Load the images into a Matrix
Mat desc_mat(images.size(), images[0].rows * images[0].cols, CV_32FC1);
for (int i=0; i<images.size(); i++) {
desc_mat.row(i) = images[i].reshape(1, 1);
}
Mat average;
PCA pca(desc_mat, average, CV_PCA_DATA_AS_ROW, nEigens);
Mat data(desc_mat.rows, nEigens, CV_32FC1); //This Mat will contain all the Eigenfaces that will be used later with SVM for detection
//Project the images onto the PCA subspace
for(int i=0; i<images.size(); i++) {
Mat projectedMat(1, nEigens, CV_32FC1);
pca.project(desc_mat.row(i), projectedMat);
data.row(i) = projectedMat.row(0);
}
CvMat d1 = (CvMat)data;
CvMat d2 = (CvMat)values;
CvSVM svm;
svm.train(&d1, &d2);
svm.save("svmdata.xml");
What etarion said is correct.
To copy a column or row you always have to write:
Mat B = mat.col(i);
A.copyTo(B);
The following program shows how to perform a PCA in OpenCV. It'll show the mean image and the first three Eigenfaces. The images I used in there are available from http://www.cl.cam.ac.uk/research/dtg/attarchive/facedatabase.html:
#include "cv.h"
#include "highgui.h"
using namespace std;
using namespace cv;
Mat normalize(const Mat& src) {
Mat srcnorm;
normalize(src, srcnorm, 0, 255, NORM_MINMAX, CV_8UC1);
return srcnorm;
}
int main(int argc, char *argv[]) {
vector<Mat> db;
// load greyscale images (these are from http://www.cl.cam.ac.uk/research/dtg/attarchive/facedatabase.html)
db.push_back(imread("s1/1.pgm",0));
db.push_back(imread("s1/2.pgm",0));
db.push_back(imread("s1/3.pgm",0));
db.push_back(imread("s2/1.pgm",0));
db.push_back(imread("s2/2.pgm",0));
db.push_back(imread("s2/3.pgm",0));
db.push_back(imread("s3/1.pgm",0));
db.push_back(imread("s3/2.pgm",0));
db.push_back(imread("s3/3.pgm",0));
db.push_back(imread("s4/1.pgm",0));
db.push_back(imread("s4/2.pgm",0));
db.push_back(imread("s4/3.pgm",0));
int total = db[0].rows * db[0].cols;
// build matrix (column)
Mat mat(total, db.size(), CV_32FC1);
for(int i = 0; i < db.size(); i++) {
Mat X = mat.col(i);
db[i].reshape(1, total).col(0).convertTo(X, CV_32FC1, 1/255.);
}
// Change to the number of principal components you want:
int numPrincipalComponents = 12;
// Do the PCA:
PCA pca(mat, Mat(), CV_PCA_DATA_AS_COL, numPrincipalComponents);
// Create the Windows:
namedWindow("avg", 1);
namedWindow("pc1", 1);
namedWindow("pc2", 1);
namedWindow("pc3", 1);
// Mean face:
imshow("avg", pca.mean.reshape(1, db[0].rows));
// First three eigenfaces:
imshow("pc1", normalize(pca.eigenvectors.row(0)).reshape(1, db[0].rows));
imshow("pc2", normalize(pca.eigenvectors.row(1)).reshape(1, db[0].rows));
imshow("pc3", normalize(pca.eigenvectors.row(2)).reshape(1, db[0].rows));
// Show the windows:
waitKey(0);
}
and if you want to build the matrix by row (like in your original question above) use this instead:
// build matrix
Mat mat(db.size(), total, CV_32FC1);
for(int i = 0; i < db.size(); i++) {
Mat X = mat.row(i);
db[i].reshape(1, 1).row(0).convertTo(X, CV_32FC1, 1/255.);
}
and set the flag in the PCA to:
CV_PCA_DATA_AS_ROW
Regarding machine learning. I wrote a document on machine learning with the OpenCV C++ API that has examples for most of the classifiers, including Support Vector Machines. Maybe you can get some inspiration there: http://www.bytefish.de/pdf/machinelearning.pdf.
data.row(i) = projectedMat.row(0);
This will not work. operator= is a shallow copy, meaning no data is actually copied. Use
cv::Mat sample = data.row(i); // also a shallow copy, points to old data!
projectedMat.row(0).copyTo(sample);
The same also for:
desc_mat.row(i) = images[i].reshape(1, 1);
I would suggest looking at the newly checked in tests in svn head
modules/core/test/test_mat.cpp
online here : https://code.ros.org/svn/opencv/trunk/opencv/modules/core/test/test_mat.cpp
has examples for PCA in the old c and new c++
Hope that helps!
Related
I have this project where we are trying to make an autonomous vehicle using a lidar and a stereo camera. To to this we're making two maps with cartographer and merging them together. However, the data from the stereo camera is not very accurate and we therefor have to manipulate the map made by cartographer. To make the camera map we are detecting lines, reading the distance and turning this into a laser scan which is the sent to cartographer. Ideally we would be able to convert the map into just the lines. This is what the camera map looks like: Camera map
What I would like to do first is fill out the holes in the map to make it easier to find lines and such later. This is where I am struggling. I have written code to convert from nav_msgs::OccupancyGrid to cv::Mat and back in addition to merging the maps. I have looked over this code and I don't think this is where the problem is. I have tried different suggestions online but have not gotten close to a solution. This is my code:
cv::Mat fill_cam_mat(cv::Mat mat) {
int thresh = 50;
cv::Mat canny_output;
cv::Canny( mat, canny_output, thresh, thresh*2 );
//std::vector<cv::Vec4i> hierarchy;
cv::Mat mat_floodfill = canny_output.clone();
cv::floodFill(mat_floodfill, cv::Point(0,0), cv::Scalar(255));
cv::Mat mat_floodfill_inv;
cv::bitwise_not(mat_floodfill, mat_floodfill_inv);
cv::Mat mat_out = (canny_output | mat_floodfill_inv);
return mat_out;
}
And my result is as follows when merged with the lidar map:
Final map
I have also tried:
cv::Mat fill_cam_mat(cv::Mat mat) {
int mat_height = mat.rows;
int mat_width = mat.cols;
int thresh = 50;
cv::Mat canny_output;
cv::Canny( mat, canny_output, thresh, thresh*2 );
cv::Mat non_zero;
cv::findNonZero(canny_output, non_zero);
std::vector<std::vector<cv::Point>> hull(non_zero.total());
for(unsigned int i = 0, n = non_zero.total(); i < n; ++i) {
cv::convexHull(non_zero, hull[i], false);
}
cv::Mat fill_contours_result(mat_height, mat_width, CV_8UC3, cv::Scalar(0));
cv::fillPoly(fill_contours_result, hull, 255);
return fill_contours_result;
}
Which gives the same result. I have also tried using cv::findContours to spicify the hull, but that worked even worse.
I am new with OpenCV and I don't understand what is wrong with my output. Would really appreciate any help on the code or if anybody have any better suggestions on how to solve the problem. Is it even necessary to fill the holes in order to get useful information from the map?
Thank you in advance!
I'm beginner in opencv. I have not gotten main concepts of opencv in details.
So maybe my code it's too dumb;
Out of my curiosity I want to try machine learning functions like a KNN, ANN.
I have the set of images with size 28*28 pixels. I want to do train cassifier for digit recognition. So first I need to assemble train set and train classes;
Mat train_data = Mat(rows, cols, CV_32FC1);
Mat train_classes = Mat(rows, 1, CV_32SC1);
Mat img = imread(image);
Mat float_data(1, cols, CV_32FC1);
img.convertTo(float_data, CV_32FC1);
How to fill train_data with float_data ?
I thought It was smth like this:
for (int i = 0; i < train_data.rows; ++i)
{
... // image is a string which contains next image path
image = imread(image);
img.convertTo(float_data, CV_32FC1);
for( int x = 0; x < train_data.cols; x++ ){
train_data.at<float> (i, x) = float_data.at<float>( x);;
}
}
KNearest knn;
knn.train(train_data, train_classes);
but it's of course doesn't work . . .
Please, tell me how to do it right. Or at least suggest the books for dummies :)
Mat train_data; // initially empty
Mat train_labels; // empty, too.
// for each img in the train set :
Mat img = imread("image_path");
Mat float_data;
img.convertTo(float_data, CV_32FC1); // to float
train_data.push_back( float_data.reshape(1,1) ); // add 1 row (flattened image)
train_labels.push_back( label_for_image ); // add 1 item
KNearest knn;
knn.train(train_data, train_labels);
it's all the same for other ml algos !
I'm trying to get the the difference between two cv::Mat frames in OpenCv. So here is what I tried,
#include<opencv2\opencv.hpp>
#include<opencv2\calib3d\calib3d.hpp>
#include<opencv2\core\core.hpp>
#include <opencv2\highgui\highgui.hpp>
int main ()
{
cv::VideoCapture cap(0);
cv::Mat frame, frame1,frame2;
int key=0;
while(key!=27){
cap >> frame;
if(key=='c'){
frame1 = frame;
key = 0;
}
if(key =='x'){
cv::absdiff(frame, frame1, frame2); // I also tried frame2= (frame -frame1)*255;
cv::imshow("difference ",frame2);
key =0;
}
cv::imshow("stream",frame);
key = cv::waitKey(10);
}
}
the result is always the same a 0 Matrix, any idea what I'm doing wrong here?
thanks in advance for your help.
Mat objects are pointer typed. After setting frame1 to frame directly using frame1 = frame, both matrices show the same point and same frame also. You have to copy frame value using "copyTo" method of Mat.
OpenCV Matrixes use pointers internally
The documentation of the Mat type states:
Mat is basically a class with two data parts: the matrix header and a pointer to the matrix containing the pixel values.
[...]
Whenever somebody copies a header of a Mat object, a counter is increased for the matrix. Whenever a header is cleaned this counter is decreased. When the counter reaches zero the matrix too is freed. Sometimes you will want to copy the matrix itself too, so OpenCV provides the clone() and copyTo() functions.
cv::Mat F = A.clone();
cv::Mat G;
A.copyTo(G);
OpenCV overloads the affectation operator on cv::Mat objects so that the line mat1 = mat2 only affects the pointer to the data in mat1 (that points to the same data as mat2). This avoids time consuming copies of all the image data.
If you want to save the data of a matrix, you have to write mat1 = mat2.clone() or mat2.copyTo(mat1).
I was looking for a similar program and I came across your post, here is a sample I have written for frameDifferencing, hope this helps, the below function will give you the difference between two frames
/** #function differenceFrame */
Mat differenceFrame( Mat prev_frame, Mat curr_frame )
{
Mat image = prev_frame.clone();
printf("frame rows %d Cols %d\n" , image.rows, image.cols);
for (int rows = 0; rows < image.rows; rows++)
{
for (int cols = 0; cols < image.cols; cols++)
{
/* printf("BGR value %lf %lf %lf\n" , abs(prev_frame.at<cv::Vec3b>(rows,cols)[0] -
curr_frame.at<cv::Vec3b>(rows,cols)[0]),
abs(prev_frame.at<cv::Vec3b>(rows,cols)[1] -
curr_frame.at<cv::Vec3b>(rows,cols)[0]),
abs(prev_frame.at<cv::Vec3b>(rows,cols)[2] -
curr_frame.at<cv::Vec3b>(rows,cols)[0]));
*/
image.at<cv::Vec3b>(rows,cols)[0] = abs(prev_frame.at<cv::Vec3b>(rows,cols)[0] -
curr_frame.at<cv::Vec3b>(rows,cols)[0]);
image.at<cv::Vec3b>(rows,cols)[1] = abs(prev_frame.at<cv::Vec3b>(rows,cols)[1] -
curr_frame.at<cv::Vec3b>(rows,cols)[1]);
image.at<cv::Vec3b>(rows,cols)[2] = abs(prev_frame.at<cv::Vec3b>(rows,cols)[2] -
curr_frame.at<cv::Vec3b>(rows,cols)[2]);
}
}
return image;
}
I'm working on a face recognition project and I am having problems when projecting on PCA subspace.
When I pass a mat vector to my funcion with the resized images, I project them, and then I reconstruct them to verify it's working well, but all I have in "Cam" window is a grey image (all same color).
I don't know what I am doing bad.
This is the function:
void doPCA (const vector<Mat>& images)
{
int nEigens = images.size()-1;
Mat data (images.size(), images[0].rows*images[0].cols, images[0].type() );
for (int i = 0; i < images.size(); i++)
{
Mat aux = data.row(i);
images[i].reshape(1,1).copyTo(aux);
}
PCA pca(data,Mat(),CV_PCA_DATA_AS_ROW,nEigens);
//Project images
Mat dataprojected(data.rows, nEigens, CV_32FC1) ;
for(int i=0; i<images.size(); i++)
{
pca.project(data.row(i), dataprojected.row(i));
}
//Backproject to reconstruct images
Mat datareconstructed (data.rows, data.cols, data.type());
for(int i=0; i<images.size(); i++)
{
pca.backProject (dataprojected.row(i), datareconstructed.row(i) );
}
for(int i=0; i<images.size(); i++)
{
imshow ("Cam", datareconstructed.row(i).reshape(1,images[0].rows) );
waitKey();
}
}
I think this post is a duplicate of:
PCA + SVM using C++ Syntax in OpenCV 2.2
Ah, I have found the error in your code. When you create the data matrix you do:
images[i].reshape(1,1).copyTo(aux);
You have to use convertTo to convert the data into the correct type and copy it to your data matrix:
images[i].reshape(1,1).convertTo(aux, CV_32FC1, 1/255.);
Then the normalized eigenvectors should be ok. And don't forget to to normalize the values between 0 and 255 before displaying them, you can use cv::normalize to do this, here's a simple function for turning it into grayscale:
Mat toGrayscale(const Mat& src) {
Mat srcnorm;
cv::normalize(src, srcnorm, 0, 255, NORM_MINMAX, CV_8UC1);
return srcnorm;
}
You may want to look at the example in my blog:
http://bytefish.de/blog/pca_in_opencv#simple_example
I'm implementing an approach from a research paper. Part of the approach calls for a major edge detector, which the authors describe as follows:
Obtain DC image (effectively downsample by 8 for both width and height)
Calculate Sobel gradient of DC image
Threshold Sobel gradient image (using T=120)
Morphological operations to clean up edge image
Note that this NOT Canny edge detection -- they don't bother with things like non-maximum suppression, etc. I could of course do this with Canny edge detection, but I want to implement things exactly as they are expressed in the paper.
That last step is the one I'm a bit stuck on.
Here is exactly what the authors say about it:
After obtaining the binary
edge map from the edge detection process, a binary morphological
operation is employed to remove isolated edge pixels,
which might cause false alarms during the edge detection
Here's how things are supposed to look like at the end of it all (edge blocks have been filled in black):
Here's what I have if I skip the last step:
It seems to be on the right track. So here's what happens if I do erosion for step 4:
I've tried combinations of erosion and dilation to obtain the same result as they do, but don't get anywhere close. Can anyone suggest a combination of morphological operators that will get me the desired result?
Here's the binarization output, in case anyone wants to play around with it:
And if you're really keen, here is the source code (C++):
#include <cv.h>
#include <highgui.h>
#include <stdlib.h>
#include <assert.h>
using cv::Mat;
using cv::Size;
#include <stdio.h>
#define DCTSIZE 8
#define EDGE_PX 255
/*
* Display a matrix as an image on the screen.
*/
void
show_mat(char *heading, Mat const &m)
{
Mat clone = m.clone();
Mat scaled(clone.size(), CV_8UC1);
convertScaleAbs(clone, scaled);
IplImage ipl = scaled;
cvNamedWindow(heading, CV_WINDOW_AUTOSIZE);
cvShowImage(heading, &ipl);
cvWaitKey(0);
}
/*
* Get the DC components of the specified matrix as an image.
*/
Mat
get_dc(Mat const &m)
{
Size s = m.size();
assert(s.width % DCTSIZE == 0);
assert(s.height % DCTSIZE == 0);
Size dc_size = Size(s.height/DCTSIZE, s.width/DCTSIZE);
Mat dc(dc_size, CV_32FC1);
cv::resize(m, dc, dc_size, 0, 0, cv::INTER_AREA);
return dc;
}
/*
* Detect the edges:
*
* Sobel operator
* Thresholding
* Morphological operations
*/
Mat
detect_edges(Mat const &src, int T)
{
Mat sobelx = Mat(src.size(), CV_32FC1);
Mat sobely = Mat(src.size(), CV_32FC1);
Mat sobel_sum = Mat(src.size(), CV_32FC1);
cv::Sobel(src, sobelx, CV_32F, 1, 0, 3, 0.5);
cv::Sobel(src, sobely, CV_32F, 0, 1, 3, 0.5);
cv::add(cv::abs(sobelx), cv::abs(sobely), sobel_sum);
Mat binarized = src.clone();
cv::threshold(sobel_sum, binarized, T, EDGE_PX, cv::THRESH_BINARY);
cv::imwrite("binarized.png", binarized);
//
// TODO: this is the part I'm having problems with.
//
#if 0
//
// Try a 3x3 cross structuring element.
//
Mat elt(3,3, CV_8UC1);
elt.at<uchar>(0, 1) = 0;
elt.at<uchar>(1, 0) = 0;
elt.at<uchar>(1, 1) = 0;
elt.at<uchar>(1, 2) = 0;
elt.at<uchar>(2, 1) = 0;
#endif
Mat dilated = binarized.clone();
//cv::dilate(binarized, dilated, Mat());
cv::imwrite("dilated.png", dilated);
Mat eroded = dilated.clone();
cv::erode(dilated, eroded, Mat());
cv::imwrite("eroded.png", eroded);
return eroded;
}
/*
* Black out the blocks in the image that contain DC edges.
*/
void
censure_edge_blocks(Mat &orig, Mat const &edges)
{
Size s = edges.size();
for (int i = 0; i < s.height; ++i)
for (int j = 0; j < s.width; ++j)
{
if (edges.at<float>(i, j) != EDGE_PX)
continue;
int row = i*DCTSIZE;
int col = j*DCTSIZE;
for (int m = 0; m < DCTSIZE; ++m)
for (int n = 0; n < DCTSIZE; ++n)
orig.at<uchar>(row + m, col + n) = 0;
}
}
/*
* Load the image and return the first channel.
*/
Mat
load_grayscale(char *filename)
{
Mat orig = cv::imread(filename);
std::vector<Mat> channels(orig.channels());
cv::split(orig, channels);
Mat grey = channels[0];
return grey;
}
int
main(int argc, char **argv)
{
assert(argc == 3);
int bin_thres = atoi(argv[2]);
Mat orig = load_grayscale(argv[1]);
//show_mat("orig", orig);
Mat dc = get_dc(orig);
cv::imwrite("dc.png", dc);
Mat dc_edges = detect_edges(dc, bin_thres);
cv::imwrite("dc_edges.png", dc_edges);
censure_edge_blocks(orig, dc_edges);
show_mat("censured", orig);
cv::imwrite("censured.png", orig);
return 0;
}
I can't imagine any combination of morphological operations that would produce the same edges as detected by the supposedly correct result, given your partial result as input.
I note that the underlying image is different; this probably contributes to why your results are so different. The Lena image is fine for indicating the type of result but not for comparisons. Do you have the exact same image as the original authors ?
What the authors described could be implemented with connected component analysis, using 8way connectivity. I would not call that morphological though.
I do think you are missing something else: Their image does not have edges that are thicker than one pixel. Yours has. The paragraph you quoted only talks about removing isolated pixels, so there must be a step you missed or implemented differently.
Good luck!
I think that what you need is a kind of erode or open that is, in a sense, 4-way and not 8-way. The default morphological kernel for OpenCV is a 3x3 rectangle (IplConvKernel with shape=CV_SHAPE_RECT). This is pretty harsh on thin edges.
You might want to try eroding with a 3x3 custom IplConvKernel with shape=CV_SHAPE_CROSS.
If you need an even finer filter, you may want to try eroding with 4 different CV_SHAPE_RECT kernels of size 1x2, 2x1 with the anchor in (0,1) and (1,0) for each.
First of all, your input image has a much higher resolution that the test input image, which can explain the fact less edges are detected - the changes are more smooth.
Second of all, since the edges are thresholded to 0, try dilation on smaller neighborhoods (e.g. compare each pixels with 4 original neighbors (in a non-serial manner)) to get rid of isolated edges.