How to determine if a cv::Mat is a zero matrix?

Question

I have a matrix that is dynamically being changed according to the following code;

 for( It=all_frames.begin(); It != all_frames.end(); ++It)
{
    ItTemp = *It;

    subtract(ItTemp, Base, NewData);

    cout << "The size of the new data for ";
    cout << " is \n" << NewData.rows << "x" << NewData.cols << endl;
    cout << "The New Data is: \n" << NewData << endl << endl;

    NewData_Vector.push_back(NewData.clone());

}

What I want to do is determine the frames at which the cv::Mat NewData is a zero matrix. I've tried comparing it to a zero matrix that is of the same size, using both the cv::compare() function and simple operators (i.e NewData == NoData), but I can't even compile the program.

Is there a simple way of determining when a cv::Mat is populated by zeroes?

Answer 1

I used

if (countNonZero(NewData) < 1) 
{
    cout << "Eye contact occurs in this frame" << endl;
}

This is a pretty simple (if perhaps not the most elegant) way of doing it.

Answer 2

To check the mat if is empty, use empty(), if NewData is a cv::Mat, NewData.empty() returns true if there's no element in NewData.

To check if it's all zero, simply, NewData == Mat::zeros(NewData.size(), NewData.type()).

Update:

After checking the OpenCV source code, you can actually do NewData == 0 to check all element is equal to 0.

Answer 3

countNonZero(Mat ) will give u number of non zeros in mat

Answer 4

How about this..

Mat img = Mat::zeros(cvSize(1024, 1024), CV_8UC3);
bool flag = true;

MatConstIterator_<double> it = img.begin<double>();
MatConstIterator_<double> it_end = img.end<double>();
for(; it != it_end; ++it)
{
    if(*it != 0)
    {
        flag = false;
        break;
    }
}

Answer 5

The Mat object has an empty property, so you can just ask Mat to tell you if it has something or it's empty. The result will be either true or false.