programnya............
#include "visualcontrol.h"
#include "webcamwidget.h"
#include <boost/bind.hpp>
#include <boost/filesystem.hpp>
#include <QDebug>
#include <QCloseEvent>
#include <QVBoxLayout>
#include <QHBoxLayout>
#include <QListWidget>
#include <QPushButton>
#include <QMessageBox>
#include <QInputDialog>
#include <logger.h>
#include "cognition/detector/facedetector.h"
#include "cognition/trainerimage.h"
namespace gui
{
const cv::Size VisualControl::testImageSize(150,200);
VisualControl::VisualControl(QWidget *parent, Qt::WFlags flags)
: QMainWindow(parent, flags)
{
setupFramework();
setupGUI();
setMinimumSize(QSize(640,480));
frameCaptureThread = boost::thread(
boost::bind(&cognition::FrameCapture::startCapturing, frameCapture));
faceDetectorThread = boost::thread(
boost::bind(&cognition::Detector::threadStart, faceDetector));
Logger::getInstance().log("Ready...");
}
VisualControl::~VisualControl()
{
}
void VisualControl::setupFramework()
{
using boost::shared_ptr;
using cognition::DetailedFaceDetector;
using cognition::EigenfaceRecognizer;
//videoCapture = shared_ptr<VideoCapture>( new VideoCapture(0) );
frameCapture = shared_ptr<cognition::FrameCapture>( new cognition::FrameCapture(32) );
//update the path of the classifier to the local path of your openCV installation!
faceDetector = shared_ptr<DetailedFaceDetector>(
new DetailedFaceDetector(DetailedFaceDetector::ALL,
"haarcascades/haarcascade_frontalface_alt.xml",
frameCapture.get(), false, 1.16));
faceDetector->loadCascade(DetailedFaceDetector::EYES, "haarcascades/haarcascade_eye.xml");
faceDetector->loadCascade(DetailedFaceDetector::NOSE, "haarcascades/haarcascade_mcs_nose.xml");
faceDetector->loadCascade(DetailedFaceDetector::MOUTH, "haarcascades/haarcascade_mcs_mouth.xml");
frameCapture->addFrameReceiver(faceDetector.get());
recognizer = shared_ptr<EigenfaceRecognizer>( new cognition::EigenfaceRecognizer );
}
void VisualControl::closeEvent(QCloseEvent *event)
{
//webcamController->unregister();
faceDetector->removeController(webcamWidget);
//qDebug() << "controller count = " << faceDetector->getControllerCount();
frameCapture->stopCapturing();
faceDetector->requestTreadStop();
frameCaptureThread.interrupt();
faceDetectorThread.interrupt();
// qDebug() << "Close event called, joining";
frameCaptureThread.join();
faceDetectorThread.join();
//event->accept();
//QMainWindow::closeEvent(event);
}
void VisualControl::setupGUI()
{
webcamWidget = new WebcamWidget(this, frameCapture.get());
faceDetector->addController(webcamWidget);
logWidget = new QListWidget;
captureTrainingImageButton = new QPushButton(tr("Capture face and store it as training image"));
connect(captureTrainingImageButton, SIGNAL(clicked()), this, SLOT(captureTrainingImage()));
trainRecognizerButton = new QPushButton(tr("Train recognizer"));
connect(trainRecognizerButton, SIGNAL(clicked()), this, SLOT(trainRecognizer()));
recognizeFaceButton = new QPushButton(tr("Recognize Visible Faces"));
connect(recognizeFaceButton, SIGNAL(clicked()), this, SLOT(recognizeFaces()));
QWidget *centralWidget = new QWidget;
QVBoxLayout *mainLayout = new QVBoxLayout;
QHBoxLayout *buttonsLayout = new QHBoxLayout;
buttonsLayout->addWidget(captureTrainingImageButton);
buttonsLayout->addWidget(trainRecognizerButton);
buttonsLayout->addWidget(recognizeFaceButton);
mainLayout->addWidget(webcamWidget);
mainLayout->addLayout(buttonsLayout);
mainLayout->addWidget(logWidget);
centralWidget->setLayout(mainLayout);
Logger::getInstance().setLogWidget(logWidget);
setCentralWidget(centralWidget);
}
void VisualControl::captureTrainingImage()
{
cv::Mat frame;
cognition::Detector::RectVector faces;
captureFrameAndFaces(faces, frame);
if(faces.size() == 0)
{
QMessageBox::information(this, "No faces found", "The detector did not find any faces!");
}
else
{
frame = frame.clone();
cognition::TrainerImage convertor(testImageSize, true, "images/");
cv::Mat faceRegion;
for(std::vector<cv::Rect>::iterator face = faces.begin(); face != faces.end(); ++face)
{
faceRegion = frame(*face);
QString filename = QInputDialog::getText(this,
tr("Image name"),
tr("enter image name (enter extension too, it determines the image format!)"));
if(filename.size() < 1) continue;
if(!convertor.processAndSaveImage(faceRegion, filename.toStdString()))
QMessageBox::information(this, "Failed", "Could not process and save the image!");
}
}
}
void VisualControl::trainRecognizer()
{
using namespace boost::filesystem;
path dir("images");
directory_iterator end;
for(directory_iterator file(dir); file != end; ++file)
{
if(is_regular_file(*file))
recognizer->addTrainingImage(file->path().string(), file->filename());
}
if(recognizer->train())
QMessageBox::information(this, "Success", "The recognizer has succesfully finished training!");
else
QMessageBox::information(this, "Error", "The recognizer has indicated that it did not train correctly!");
}
void VisualControl::recognizeFaces()
{
if(recognizer->trained())
{
cv::Mat frame;
cognition::Detector::RectVector faces;
captureFrameAndFaces(faces, frame);
if(faces.size() == 0)
QMessageBox::information(this, "No faces found", "The detector did not find any faces!");
else
{
frame = frame.clone();
cognition::TrainerImage convertor(testImageSize, true);
cv::Mat faceRegion;
cv::Mat processedFace;
for(std::vector<cv::Rect>::iterator face = faces.begin(); face != faces.end(); ++face)
{
faceRegion = frame(*face);
processedFace = convertor.processImage(faceRegion);
std::string name = recognizer->recognize(processedFace);
Logger::getInstance().log(std::string("Recognized: ") + name);
}
}
}
else
{
QMessageBox::information(this, "Recognizer is not trained", "Recognizer is not trained or failed to train, add enough training images and train the recognizer!");
}
}
void VisualControl::captureFrameAndFaces(cognition::Detector::RectVector &rects, cv::Mat &frame)
{
rects = webcamWidget->getCurrentFaces();
frame = webcamWidget->getCurrentFrame();
//cognition::FaceDetector detector("C:/OpenCV2.1/data/haarcascades/haarcascade_frontalface_alt.xml");
//*frameCapture->getCaptureDevice() >> frame;
//detector.receiveFrame(frame);
//detector.processFrame();
//rects = detector.getAreas();
}
}
