stiv-yakovenko's profile - activity

invalid use of incomplete type ?struct cv::DefaultDeleter<cvhaarclassifiercascade> I am trying to build this c++ s

invalid use of incomplete type ?struct cv::DefaultDeleter<cvhaarclassifiercascade> I am trying to build this c++ s

thank you dear friend!

I am running this simple application to perform image resize on GTX1080ti GPU:

#include <opencv2/opencv.hpp>
#include "opencv2/cudaimgproc.hpp"
#include "opencv2/cudawarping.hpp"
using namespace std;
using namespace cv;
using namespace cv::cuda;
static void gpuResize(Mat in, Mat out){
    double k = in.cols/416.;
    cuda::GpuMat gpuInImage;
    cuda::GpuMat gpuOutImage;
    gpuInImage.upload(in);
    const Size2i &newSize = Size(416, in.rows / k);
    cout << "newSize " << newSize<< endl;
    cuda::resize(gpuInImage, gpuOutImage, newSize);
    gpuOutImage.download(out);
}

int main(){
    Mat im = Mat::zeros(Size(832,832),CV_8UC3);
    Mat out;
    if (getCudaEnabledDeviceCount() == 0){
        return cerr << "No GPU found or the library is compiled without CUDA support" << endl, -1;
    }
    cv::cuda::printShortCudaDeviceInfo(cv::cuda::getDevice());
    gpuResize(im,out);
    cout << "real size="<<out.size() <<endl;
}

the output I'm getting:

Device 0:  "GeForce GTX 1080 Ti"  11177Mb, sm_61, Driver/Runtime ver.10.10/10.10
newSize [416 x 416]
real size=[0 x 0]

why output matrix has 0x0 size, how should I fix that?

To use SIFT you have to downgrade to 3.4.2.16: more details here

How to resize image with nvidia GPU? I am running this simple application to perform image resize on GTX1080ti GPU: #in

it gives me: no element "v4l2video1h264enc"

I confirm same behavior as of today (OPENCL + jetson nano)

undefined reference to cv::cuda::dft I've build opencv with CUDA support on my CentOS linux with NVIDIA with CUDA suppor

I also tried using samples from cv::text and they all work stably only on example data

Poor performance of cv::text I need to locate all letters on this screenshot: Tried https://github.com/opencv/opencv_

1st link is a deadlink

@StevePuttemans, CAP_PROP_FRAME_COUNT is unreliable, for some files it gives huge negative value and iterating frames is

Going through each frame with VideoCapture might be slow. You can get 3x perfomance boost if you invoke ffprobe.exe, ass

I have same problem with debug build made with vcpkg.

stereoRectify doesn't work sometimes I have many shots of object on aruco board: All images in the folder are undisto

How can I convert cordinates from rectified stereo to original coordinates in opencv? I am using this code to rectify st

To load camera calibration matrix like this: %YAML:1.0 --- image_width: 4000 image_height: 3000 camera_matrix: !!opencv

Output from solvePnP doesn't match projectPoints I get strange data from solvePnP, so I tried to check it with projectPo

for opencv 3 it should be not path-to-build-directory-of-VTK, but vtk cmake file folder

Sorry, this doesn't help me on windows, with vtk and successfull cmake for it, there still no opencv_viz.dll after build