ranger's profile - activity

Include all the libraries in directory C:\opencv\build\x86\mingw\bin into the project file...

Sir I have a static background...I need something like this https://www.youtube.com/watch?v=KRKKektCcok sir... Cant really find it but I m trying...I m very new to image processing...

Okay I have ezperimented with lot of haar cascades, HOG and all but none of them was very efficient to meet my purpose. I am working on a video processing project which include tracking of humans (mainly). I have a static background in hand (always). So I thought detecting the foreground objects would be very easy. I used subtracting method like this :

cv::Mat temp;
cv::subtract(camFrame,backImage,temp);
cv::threshold(temp,temp,25,255,cv::THRESH_BINARY);

But this doesnt seem to work well in getting only the foreground objects.I thk the varying lighting effects is also given as foreground from this. Is there anything I am missing in this process??. Everybody said its very easy if theres a static background. I am very new to image processing in opencv. I will be glad if somebody can help me out and appreciate some code as well

0 down vote favorite

I am working on a video processing project which involves tracking objects which can be stationary or moving. The actual static background is available at the beginning. I have writeen a simple code for getting the foreground :

cv::Mat temp;
cv::absdiff(background,camFrame,temp);
threshold(temp, temp, 128, 255, CV_THRESH_BINARY | CV_THRESH_OTSU);
cv::erode(temp,temp,cv::Mat());

But unfortunately the code doesnt seem to work well in retrieving the foreground objects...I will be glab if someboday provide me with some suggestions and bits of code...

images and saves it in a folder with "label-number" format, so that i can later use these images to train the fisherface recognizer model. But when I execute the model->train statement its showin the above defined problem....I am wokring on it for a week..couldn't find a solution... I have uploaded the full directory of my code here : http://www.4shared.com/rar/ZQmAyz7K/QtTracker3.html

The exe in the debug folder doesnt seem to work directly on double click...but it executes when compiled and run inside Qt creator IDE...At first you need to ADD new face and using AUTO / TAKE SHOT in the child window you can save the images in grayscale as database...Ten once close the child window and exceute TRAIN in parent...then the error pops out...:-((((

thanks a lot sir, atleast somebody commented.... the full code is available here https://github.com/kocchumon/qttracker . I compiled the code using mingw 4.4 using qt creator IDE in Windows 7 32 bit edition....and the when I run the code windows says :

Problem signature: Problem Event Name: APPCRASH Application Name: QtTracker3.exe Application Version: 0.0.0.0 Application Timestamp: 5105345b Fault Module Name: libopencv_core242.dll Fault Module Version: 0.0.0.0 Fault Module Timestamp: 50da6896 Exception Code: c0000005 Exception Offset: 000ac3f6 OS Version: 6.1.7600.2.3.0.256.1 Locale ID: 1033

and Qt creator says - QtTracker3.exe exited with code -1073741819 Now about the program....the program is able to track human face and takes series of gray scale ......

I am working on a video processing and the preliminary steps include face recognition...I made a data base of the faces (100*100 and grayscale) and stored them in mat vector and corresponding labels in int vector but while I call the train function for the model it gives an error -1073741819 full code can be found here

If you really want to detect and track human eye I would suggest you to detect the human face first and track it and simultaneously detect the eye...I thk this can really reduce the window to be searched and efficient if the succeeding search for eye is limited within the window...DOnt forget to update the window after processing every frame...

I am working on a video processing project and the priliminary step involves face recognition. As i was unable to train fisher/ eigen face recognizer models I tried using LBP face recognizer model and it jst worked...Basically what my face recognition program does is it jst draws a rectangle around the recognized face and I dont want any rectangle around alien faces...but LPB recognizer forces to do so?...because it predicts the nearest label the detected face matches....got any idea guys to limit the recognition to the nearest..

Have you installed android sdk for windows??

I am working on a project based on face recognition and modefied the code here such that the image file paths are read using Qt directory libraries and is working fine. Problem occurs when I try to train the fisher faces model ,i.e the model->train() call let the app to crash...and Qt creator says the program exited with code -1073741819

void Dialog::on_btnAdd_2_clicked()
{
    tmrTimer->stop();

    dir=QDir(directory);
    QFileInfoList files;
    QFileInfo file;
    QString filen;

    char name[10];
    int i,j;
    std::string str;
    std::string filename;

    dir.setFilter(QDir::Files | QDir::Hidden | QDir::NoSymLinks);
    files=dir.entryInfoList();
    int nofiles;

    nofiles=files.size();
    for(i=0;i<nofiles;i++)
        {
            file=files.at(i);
            filen=file.baseName();
            filename=filen.toStdString();
            strcpy(name,filename.c_str());

            for(j=0;name[j]!='-';j++);
            name[j]='\0';                                   //label value ends with '-' in image names

            filen=file.absoluteFilePath();
            str=filen.toStdString();

            /*
            cv::Mat offimage=cv::imread(str.c_str(),-1);

            cv::namedWindow("face",WINDOW_AUTOSIZE);        //show retrieved image in a window
            cv::imshow("face",offimage);
            cvvWaitKey(1000);
           */

            newImages.push_back(cv::imread(str.c_str(),CV_LOAD_IMAGE_GRAYSCALE));
            newLabels.push_back(atoi(name));


        }


    ui->boxConsole->appendPlainText("Training Started....!!");

    cv::Ptr<cv::FaceRecognizer> model = cv::createFisherFaceRecognizer();
    model->train(newImages,newLabels);       //this causes the error

    //strcat(home,"\\data.xml");

    //model->save(home);

    ui->boxConsole->appendPlainText("Training Completed!!");
    tmrTimer->start();
}

I am using windows 7 environment and ide is Qt creator. The crash report by the OS says that :

Problem signature:
  Problem Event Name:   APPCRASH
  Application Name: QtTracker3.exe
  Application Version:  0.0.0.0
  Application Timestamp:    51035d17
  Fault Module Name:    libopencv_core242.dll
  Fault Module Version: 0.0.0.0
  Fault Module Timestamp:   50da6896
  Exception Code:   c0000005
  Exception Offset: 000a38f0
  OS Version:   6.1.7600.2.3.0.256.1
  Locale ID:    1033
  Additional Information 1: 0a9e
  Additional Information 2: 0a9e372d3b4ad19135b953a78882e789
  Additional Information 3: 0a9e
  Additional Information 4: 0a9e372d3b4ad19135b953a78882e789

full code can be found here