GTK+ moving window based on face detection coordinates - c++

Question

I have a program here that performs face detection and I would like to use these coordinates to move a window created with GTK+ 3.22 using GTK's
gtk_window_move function. I would like the window to remain open the entire time while it is moving similar to OpenCV's moveWindow function.
I just downloaded the GTK+ packages yesterday so I am not all too familiar.
The program will perform a loop 100 times, tracking a face the entire time. Currently, the face tracking works, but the window does not appear until the loop is complete. Why is this? I believe the gtk_move_window function is working, but the window does not stay open. I have tried reopening the window each time in the loop, or just opening once before the loop. If you are familiar with OpenCV's moveWindow function that is exactly what I am looking for. Here is the sample code.
By the way, if you know how a GTK+ function that will bring the window to the very top layer on top of all the other windows open when called, that is helpful information for me as well.
#include "FlyCapture2.h"
#include <opencv2/core/core.hpp>
#include <opencv2/objdetect/objdetect.hpp>
#include <opencv2/imgproc/imgproc.hpp>
#include <opencv2/core/cuda.hpp>
#include <opencv2/cudaobjdetect.hpp>
#include <math.h>
#include <thread>
#include <iostream>
#include <vector>
#include <gtk-3.0/gtk/gtk.h>
using namespace FlyCapture2;
cv::Ptr<cv::cuda::CascadeClassifier> face_detect;
void detect_faces(cv::Mat img, cv::cuda::GpuMat buf,GtkWidget *win)
{
std::vector<cv::Rect>faces;
cv::cuda::GpuMat image_gpu(img);
//Face detection here
...
if (faces.size() > 0)
{
float x = faces[0].x;
float y = faces[0].y;
int new_x = roundf(x*40/51);
int new_y = roundf(y*135/256);
gtk_window_move(GTK_WINDOW (win),new_x,new_y);
gtk_widget_show (win); //Should this go here?
std::cout<<faces[0]<<std::endl;
}
}
int main( int argc, char *argv[])
{
//Camera connect here
...
//face detect variables
face_detect = cv::cuda::CascadeClassifier::create("/home/nvidia/opencv/data/haarcascades_cuda/haarcascade_frontalface_default.xml");
cv::cuda::GpuMat objbuf;
//GTK+ Params
GtkWidget *window;
GdkRGBA *color;
gtk_init (&argc, &argv);
gdk_rgba_parse(color,"(0,0,0)");
window = gtk_window_new (GTK_WINDOW_TOPLEVEL);
gtk_window_set_decorated(GTK_WINDOW (window),FALSE);
gtk_window_set_position(GTK_WINDOW (window), GTK_WIN_POS_CENTER);
gtk_widget_override_background_color(window, GTK_STATE_FLAG_NORMAL, color);
gtk_widget_show (win); //Should this go here?
// capture loop
for (int i=0;i<100;i++)
{
// Get the image
Image rawImage;
camera.RetrieveBuffer( &rawImage );
// convert to rgb
Image rgbImage;
rawImage.Convert( FlyCapture2::PIXEL_FORMAT_MONO8, &rgbImage );
// convert to OpenCV Mat
unsigned int rowBytes = (double)rgbImage.GetReceivedDataSize()/(double)rgbImage.GetRows();
cv::Mat image = cv::Mat(rgbImage.GetRows(), rgbImage.GetCols(), CV_8UC1, rgbImage.GetData(),rowBytes);
//Detect Faces
detect_faces(image,objbuf,window);
}
//Disconnect Camera
camera.StopCapture();
camera.Disconnect();
gtk_main();
return 0;
}

Answer 1

The code in your capture loop should be in an event handler callback.
You first need to call g_timeout_add or g_idle_add to register your callback.
The callback you registered is a GSourceFunc that will be called after gtk_main is run. The return value (G_SOURCE_CONTINUE or G_SOURCE_REMOVE) controls if you want to have the callback be called again.