I would like to fill a matrix column wise. I have the following numpy code which I am having difficulty converting to C++ Armadillo.
# numpy code
m = np.zeros((nrows, nrows))
# fill a matrix of lags
for i in range(0, nrows):
r = np.roll(vec_v, i)
m[:, i] = r
where vec_v is a single column vector and nrows is the number of rows in that column vector.
This is my Armadillo attempt
# armadillo conversion
mat m(nrows, nrows); m.zeroes();
for(int i = 0; i < nrows; i++){
vec r = shift(vec_v, i)
m.col(i).fill(r);
}
What is the reccommended way to initialize a matrix then fill the values column-wise.
The = operator should work here.
mat m(nrows, nrows); m.zeros();
for(int i = 0; i < nrows; i++){
vec r = shift(vec_v, i);
m.col(i) = r;
}
Matrix initialization can be simplified and the generation of the temporary r vector can be avoided, as below.
mat m(nrows, nrows, fill::zeros);
for(int i = 0; i < nrows; i++){
m.col(i) = shift(vec_v, i);
}
This question has been asked here, but still no answer/solution. My problem is this: I trained an SVM (with RBF kernel) for smoke detection, using the RGB histogram distribution (in 8 bins - so M = 24) of the smoke:
cv::Mat labelsMat = cv::Mat(N, 1, CV_32SC1);
for (int i = 0; i < N; i++)
{
labelsMat.at<int>(i, 0) = labels[i];
}
cv::Mat trainingDataMat = cv::Mat(N, M, CV_32FC1);
for (int i = 0; i < N; i++)
{
for (int j = 0; j < M; j++)
{
trainingDataMat.at<float>(i, j) = histogramData[i][j];
}
}
// Create the SVM
cv::Ptr<ml::SVM> svm = ml::SVM::create();
svm->setType(ml::SVM::C_SVC);
svm->setKernel(ml::SVM::RBF);
svm->setTermCriteria(cv::TermCriteria(cv::TermCriteria::MAX_ITER, 1000, 1e-8));
// Train the SVM
svm->trainAuto(trainingDataMat, ml::ROW_SAMPLE, labelsMat);
svm->save(SVMFileName);
Then I saved the SVM model in a file. For the detection, after loading the SVM model:
svm = cv::ml::SVM::load(SVMFile);
I proceeded with the smoke detection; in this case to decide for each detected blob in a frame whether it's smoke or not:
for (int i = 0; i < 8; i++)
histogramData.at<float>(0, i) = Rhist[i];
for (int i = 8; i < 16; i++)
histogramData.at<float>(0, i) = Ghist[i];
for (int i = 16; i < 24; i++)
histogramData.at<float>(0, i) = Bhist[i];
float response = svm->predict(histogramData);
The frames where detection (true/false positive) occurs are saved, with the frame no. When I run this on the same video several times, each time different results (frame no.) will be produced (the blob detection always produces the same blobs). Regarding the detection, sometimes (most of the time) the smoke will be detected, but there are some cases where the same smoke will not be detected (the same video).
Anybody has any idea how to resolve this? Or is this still a known problem in OpenCV SVM?
Just realized my stupid error in the code: the indexing of Ghist & Bhist to form the data for prediction is totally incorrect, hence the inconsistencies!
I have an image data set that I would like to partition into k clusters. I am trying to use the opencv implementation of k-means clustering.
Firstly, I store my Mat images into a vector of Mat and then I am trying to use the kmeans function. However, I am getting an assertion error.
Should the images be stored into a different kind of structure? I have read the k-means documentation and I dont seem to understand what I am doing wrong. This is my code:
Thank you in advance,
vector <Mat> images;
string folder = "D:\\football\\positive_clustering\\";
string mask = "*.bmp";
vector<string> files = getFileList(folder + mask);
for (int i = 0; i < files.size(); i++)
{
Mat img = imread(folder + files[i]);
images.push_back(img);
}
cout << "Vector of positive samples created" << endl;
int k = 10;
cv::Mat bestLabels;
cv::kmeans(images, k, bestLabels, TermCriteria(), 3, KMEANS_PP_CENTERS);
//have a look
vector<cv::Mat> clusterViz(bestLabels.rows);
for (int i = 0; i<bestLabels.rows; i++)
{
clusterViz[bestLabels.at<int>(i)].push_back(cv::Mat(images[bestLabels.at<int>(i)]));
}
namedWindow("clusters", WINDOW_NORMAL);
for (int i = 0; i<clusterViz.size(); i++)
{
cv::imshow("clusters", clusterViz[i]);
cv::waitKey();
}
I have two Vec3b images and I want to find the MSE (Mean Square Error) between them. I know how to do it when you have two uchar images, but when you have two Vec3b images where there are 3 different values stored for each pixel how do you calculate it?
You should compute the Euclidean distance for each pair of pixels:
MSE = 0;
for(int i = 0; i < width; i++)
for(int j = 0; j < height; j++)
MSE += sqrt(pow(img1.at<Vec3b>(j, i)[0] - img2.at<Vec3b>(j, i)[0]), 2) + pow(img1.at<Vec3b>(j, i)[1] - img2.at<Vec3b>(j, i)[1]), 2) + pow(img1.at<Vec3b>(j, i)[2] - img2.at<Vec3b>(j, i)[2]), 2));
MSE /= width * height;
This code can be optimized and if you convert your image from BGR to HSV, you could get better results according what you want to do.
To calculate the Mean Square Error for 1D and 3D images in opencv, you can use this post which might be faster since image scanning takes longer times.
double getMSE(Mat& I1, Mat& I2)
{
Mat s1;
// save the I! and I2 type before converting to float
int im1type = I1.type();
int im2type = I2.type();
// convert to float to avoid producing zero for negative numbers
I1.convertTo(I1, CV_32F);
I2.convertTo(I2, CV_32F);
absdiff(I1, I2, s1); // |I1 - I2|
s1.convertTo(s1, CV_32F); // cannot make a square on 8 bits
s1 = s1.mul(s1); // |I1 - I2|^2
Scalar s = sum(s1); // sum elements per channel
double sse = s.val[0] + s.val[1] + s.val[2]; // sum channels
if( sse <= 1e-10) // for small values return zero
return 0;
else
{
double mse =sse /(double)(I1.channels() * I1.total());
return mse;
// Instead of returning MSE, the tutorial code returned PSNR (below).
//double psnr = 10.0*log10((255*255)/mse);
//return psnr;
}
// return I1 and I2 to their initial types
I1.convertTo(I1, im1type);
I2.convertTo(I2, im2type);
}
The above code returns zero for small mse values (under 1e-10). Terms s.val1 and s.val[2] are zero for 1D images.
If you want to check for 1D image input, use the following code to test (with random unsigned numbers):
Mat I1(12, 12, CV_8UC1), I2(12, 12, CV_8UC1);
double low = 0;
double high = 255;
cv::randu(I1, Scalar(low), Scalar(high));
cv::randu(I2, Scalar(low), Scalar(high));
double mse = getMSE(I1, I2);
cout << mse << endl;
If you want to check for 3D image input, use the following code to test (with random unsigned numbers):
Mat I1(12, 12, CV_8UC3), I2(12, 12, CV_8UC3);
double low = 0;
double high = 255;
cv::randu(I1, Scalar(low), Scalar(high));
cv::randu(I2, Scalar(low), Scalar(high));
double mse = getMSE(I1, I2);
cout << mse << endl;
I'm trying to make a mobile fast version of Gaussian Blur image filter.
I've read other questions, like: Fast Gaussian blur on unsigned char image- ARM Neon Intrinsics- iOS Dev
For my purpose i need only a fixed size (7x7) fixed sigma (2) Gaussian filter.
So, before optimizing for ARM NEON, I'm implementing 1D Gaussian Kernel in C++, and comparing performance with OpenCV GaussianBlur() method directly in mobile environment (Android with NDK). This way it will result in a much simpler code to optimize.
However the result is that my implementation is 10 times slower then OpenCV4Android version. I've read that OpenCV4 Tegra have optimized GaussianBlur implementation, but I don't think that standard OpenCV4Android have those kind of optimizations, so why is my code so slow?
Here is my implementation (note: reflect101 is used for pixel reflection when applying filter near borders):
Mat myGaussianBlur(Mat src){
Mat dst(src.rows, src.cols, CV_8UC1);
Mat temp(src.rows, src.cols, CV_8UC1);
float sum, x1, y1;
// coefficients of 1D gaussian kernel with sigma = 2
double coeffs[] = {0.06475879783, 0.1209853623, 0.1760326634, 0.1994711402, 0.1760326634, 0.1209853623, 0.06475879783};
//Normalize coeffs
float coeffs_sum = 0.9230247873f;
for (int i = 0; i < 7; i++){
coeffs[i] /= coeffs_sum;
}
// filter vertically
for(int y = 0; y < src.rows; y++){
for(int x = 0; x < src.cols; x++){
sum = 0.0;
for(int i = -3; i <= 3; i++){
y1 = reflect101(src.rows, y - i);
sum += coeffs[i + 3]*src.at<uchar>(y1, x);
}
temp.at<uchar>(y,x) = sum;
}
}
// filter horizontally
for(int y = 0; y < src.rows; y++){
for(int x = 0; x < src.cols; x++){
sum = 0.0;
for(int i = -3; i <= 3; i++){
x1 = reflect101(src.rows, x - i);
sum += coeffs[i + 3]*temp.at<uchar>(y, x1);
}
dst.at<uchar>(y,x) = sum;
}
}
return dst;
}
A big part of the problem, here, is that the algorithm is overly precise, as #PaulR pointed out. It's usually best to keep your coefficient table no more precise than your data. In this case, since you appear to be processing uchar data, you would use roughly an 8-bit coefficient table.
Keeping these weights small will particularly matter in your NEON implementation because the narrower you have the arithmetic, the more lanes you can process at once.
Beyond that, the first major slowdown that stands out is that having the image edge reflection code within the main loop. That's going to make the bulk of the work less efficient because it will generally not need to do anything special in that case.
It might work out better if you use a special version of the loop near the edges, and then when you're safe from that you use a simplified inner loop that doesn't call that reflect101() function.
Second (more relevant to prototype code) is that it's possible to add the wings of the window together before applying the weighting function, because the table contains the same coefficients on both sides.
sum = src.at<uchar>(y1, x) * coeffs[3];
for(int i = -3; i < 0; i++) {
int tmp = src.at<uchar>(y + i, x) + src.at<uchar>(y - i, x);
sum += coeffs[i + 3] * tmp;
}
This saves you six multiplies per pixel, and it's a step towards some other optimisations around controlling overflow conditions.
Then there are a couple of other problems related to the memory system.
The two-pass approach is good in principle, because it saves you from performing a lot of recomputation. Unfortunately it can push the useful data out of L1 cache, which can make everything quite a lot slower. It also means that when you write the result out to memory, you're quantising the intermediate sum, which can reduce precision.
When you convert this code to NEON, one of the things you will want to focus on is trying to keep your working set inside the register file, but without discarding calculations before they've been fully utilised.
When people do use two passes, it's usual for the intermediate data to be transposed -- that is, a column of input becomes a row of output.
This is because the CPU will really not like fetching small amounts of data across multiple lines of the input image. It works out much more efficient (because of the way the cache works) if you collect together a bunch of horizontal pixels, and filter those. If the temporary buffer is transposed, then the second pass also collects together a bunch of horizontal points (which would vertical in the original orientation) and it transposes its output again so it comes out the right way.
If you optimise to keep your working set localised, then you might not need this transposition trick, but it's worth knowing about so that you can set yourself a healthy baseline performance. Unfortunately, localisation like this does force you to go back to the non-optimal memory fetches, but with the wider data types that penalty can be mitigated.
If this is specifically for 8 bit images then you really don't want floating point coefficients, especially not double precision. Also you don't want to use floats for x1, y1. You should just use integers for coordinates and you can use fixed point (i.e. integer) for the coefficients to keep all the filter arithmetic in the integer domain, e.g.
Mat myGaussianBlur(Mat src){
Mat dst(src.rows, src.cols, CV_8UC1);
Mat temp(src.rows, src.cols, CV_16UC1); // <<<
int sum, x1, y1; // <<<
// coefficients of 1D gaussian kernel with sigma = 2
double coeffs[] = {0.06475879783, 0.1209853623, 0.1760326634, 0.1994711402, 0.1760326634, 0.1209853623, 0.06475879783};
int coeffs_i[7] = { 0 }; // <<<
//Normalize coeffs
float coeffs_sum = 0.9230247873f;
for (int i = 0; i < 7; i++){
coeffs_i[i] = (int)(coeffs[i] / coeffs_sum * 256); // <<<
}
// filter vertically
for(int y = 0; y < src.rows; y++){
for(int x = 0; x < src.cols; x++){
sum = 0; // <<<
for(int i = -3; i <= 3; i++){
y1 = reflect101(src.rows, y - i);
sum += coeffs_i[i + 3]*src.at<uchar>(y1, x); // <<<
}
temp.at<uchar>(y,x) = sum;
}
}
// filter horizontally
for(int y = 0; y < src.rows; y++){
for(int x = 0; x < src.cols; x++){
sum = 0; // <<<
for(int i = -3; i <= 3; i++){
x1 = reflect101(src.rows, x - i);
sum += coeffs_i[i + 3]*temp.at<uchar>(y, x1); // <<<
}
dst.at<uchar>(y,x) = sum / (256 * 256); // <<<
}
}
return dst;
}
This is the code after implementing all the suggestions of #Paul R and #sh1, summarized as follows:
1) use only integer arithmetic (with precision to taste)
2) add the values of the pixels at the same distance from the mask center before applying the multiplications, to reduce the number of multiplications.
3) apply only horizontal filters to take advantage of the storage by rows of the matrices
4) separate cycles around the edges from those inside the image not to make unnecessary calls to reflection functions. I totally removed the functions of reflection, including them inside the loops along the edges.
5) In addition, as a personal observation, to improve rounding without calling a (slow) function "round" or "cvRound", I've added to both temporary and final pixel results 0.5f (= 32768 in integers precision) to reduce the error / difference compared to OpenCV.
Now the performance is much better from about 15 to about 6 times slower than OpenCV.
However, the resulting matrix is not perfectly identical to that obtained with the Gaussian Blur of OpenCV. This is not due to arithmetic length (sufficient) as well as removing the error remains. Note that this is a minimum difference, between 0 and 2 (in absolute value) of pixel intensity, between the matrices resulting from the two versions. Coefficient are the same used by OpenCV, obtained with getGaussianKernel with same size and sigma.
Mat myGaussianBlur(Mat src){
Mat dst(src.rows, src.cols, CV_8UC1);
Mat temp(src.rows, src.cols, CV_8UC1);
int sum;
int x1;
double coeffs[] = {0.070159, 0.131075, 0.190713, 0.216106, 0.190713, 0.131075, 0.070159};
int coeffs_i[7] = { 0 };
for (int i = 0; i < 7; i++){
coeffs_i[i] = (int)(coeffs[i] * 65536); //65536
}
// filter horizontally - inside the image
for(int y = 0; y < src.rows; y++){
uchar *ptr = src.ptr<uchar>(y);
for(int x = 3; x < (src.cols - 3); x++){
sum = ptr[x] * coeffs_i[3];
for(int i = -3; i < 0; i++){
int tmp = ptr[x+i] + ptr[x-i];
sum += coeffs_i[i + 3]*tmp;
}
temp.at<uchar>(y,x) = (sum + 32768) / 65536;
}
}
// filter horizontally - edges - needs reflect
for(int y = 0; y < src.rows; y++){
uchar *ptr = src.ptr<uchar>(y);
for(int x = 0; x <= 2; x++){
sum = 0;
for(int i = -3; i <= 3; i++){
x1 = x + i;
if(x1 < 0){
x1 = -x1;
}
sum += coeffs_i[i + 3]*ptr[x1];
}
temp.at<uchar>(y,x) = (sum + 32768) / 65536;
}
}
for(int y = 0; y < src.rows; y++){
uchar *ptr = src.ptr<uchar>(y);
for(int x = (src.cols - 3); x < src.cols; x++){
sum = 0;
for(int i = -3; i <= 3; i++){
x1 = x + i;
if(x1 >= src.cols){
x1 = 2*src.cols - x1 - 2;
}
sum += coeffs_i[i + 3]*ptr[x1];
}
temp.at<uchar>(y,x) = (sum + 32768) / 65536;
}
}
// transpose to apply again horizontal filter - better cache data locality
transpose(temp, temp);
// filter horizontally - inside the image
for(int y = 0; y < src.rows; y++){
uchar *ptr = temp.ptr<uchar>(y);
for(int x = 3; x < (src.cols - 3); x++){
sum = ptr[x] * coeffs_i[3];
for(int i = -3; i < 0; i++){
int tmp = ptr[x+i] + ptr[x-i];
sum += coeffs_i[i + 3]*tmp;
}
dst.at<uchar>(y,x) = (sum + 32768) / 65536;
}
}
// filter horizontally - edges - needs reflect
for(int y = 0; y < src.rows; y++){
uchar *ptr = temp.ptr<uchar>(y);
for(int x = 0; x <= 2; x++){
sum = 0;
for(int i = -3; i <= 3; i++){
x1 = x + i;
if(x1 < 0){
x1 = -x1;
}
sum += coeffs_i[i + 3]*ptr[x1];
}
dst.at<uchar>(y,x) = (sum + 32768) / 65536;
}
}
for(int y = 0; y < src.rows; y++){
uchar *ptr = temp.ptr<uchar>(y);
for(int x = (src.cols - 3); x < src.cols; x++){
sum = 0;
for(int i = -3; i <= 3; i++){
x1 = x + i;
if(x1 >= src.cols){
x1 = 2*src.cols - x1 - 2;
}
sum += coeffs_i[i + 3]*ptr[x1];
}
dst.at<uchar>(y,x) = (sum + 32768) / 65536;
}
}
transpose(dst, dst);
return dst;
}
According to Google document, on Android device, using float/double is twice slower than using int/uchar.
You may find some solutions to speed up your C++ code on this Android documents.
https://developer.android.com/training/articles/perf-tips