mhaghighat's profile - activity

You can release the old tracker and create a new tracker and initiate it with your new ROI. That is: tracker.release();

It seems it does not preserve the aspect ratio: https://github.com/opencv/opencv/blob/9e1b1e5389237c2b9f6c7b9d7715d9836c

Say, I have a

cv::Ptr<cv::backgroundsubtractormog2> mogPtr = cv::createBackgroundSubtractorMOG2(300, 16, true);

Later, I'd like to change the configuration of the background subtractor, but I want to reassign the new object pointer to the same variable (mogPtr):

mogPtr = cv::createBackgroundSubtractorMOG2(300, 64, false);

I wonder if I should release the mogPtr before I assign a new pointer to it. Is it a bad idea if I do? Will there be a memory leak if I don't?

Should I release cv::Ptr before reassigning? Say, I have a cv::Ptr<cv::backgroundsubtractormog2> mogPtr = cv:

Should I release cv::Ptr before reassigning? Say, I have a cv::Ptr<cv::backgroundsubtractormog2> mogPtr = cv:

Should I release cv::Ptr before reassigning? Say, I have a cv::Ptr<cv::backgroundsubtractormog2> mogPtr= cv::

So, it does not preserve the aspect ratio. Thanks a lot.

Does OpenCV blobFromImage preserve the aspect ratio of the input image? I wonder if the cv::blobFromImage() preserves th

What about OpenFace?

The issue is due to the missing python3.dll file in Anaconda3. To fix the issue, you should simply copy the python3.dl

CMake Build Error OpenCV 3.4.0 System information: OpenCV => 3.4.0 Operating System / Platform => Windows 7 (64

CMake Build Error OpenCV 3.4.0 System information: OpenCV => 3.4.0 Operating System / Platform => Windows 7 (64

CMake Build Error OpenCV 3.4.0 System information: OpenCV => 3.4.0 Operating System / Platform => Windows 7 (64

Assuming both image1 and image2 are the same size: float correlation(cv::Mat &image1, cv::Mat &image2) { c

Why do you have to convert the input images to float?

cv2.BORDER_REFLECT - Border will be mirror reflection of the border elements, like this : fedcba|abcdefgh|hgfedcb cv2.

You can release the old tracker and create a new tracker and initiate it with your new ROI. That is: tracker.release();

p was a nullptr!

I run a KCF tracker over the frames of a video in which an object appears from an arbitrary corner and moves towards another arbitrary corner and exits the frame region. The tracker is supposed to follow this object all along.

It works pretty well most of the times, but sometimes it crashes when the object's region of interest (ROI) falls out of the valid frame region, i.e., the rectangle between (0,0) and (frame.size().width-1, frame.size().height-1).

It works fine most of the times even when half of the ROI is out of the valid frame region, and it keeps tracking the object when it returns to the valid region. However, sometimes, it crashes when the ROI goes over the invalid region. Since the crash happens rarely, I don't want to impose conditions to terminate the tracker as soon as the ROI touches an invalid region, but I don't want the crash either.

I couldn't find any documentation stating the termination rule of the cv::Tracker. I wonder if anyone knows what is an acceptable ROI and what is not!e ROI and what is not!

Try this:

cv::Ptr<cv::BackgroundSubtractorMOG2> pMOG2 = cv::createBackgroundSubtractorMOG2();
cv::Mat fgMask;  // Foreground mask

// Update the background model and get the foreground mask
pMOG2->apply(frame, fgMask);

// Blur the foreground mask to reduce the effect of noise and false positives
cv::blur(fgMask, fgMask, cv::Size(15, 15), cv::Point(-1, -1));

// Remove the shadow parts and the noise
cv::threshold(fgMask, fgMask, 128, 255, cv::THRESH_BINARY);

Is there a built-in function to apply a whitening transformation on a cv::Mat with 3 channels? I want the image to have a zero mean and a unit variance.

I am having a similar problem, it tries to open file 'opencv_cudaobjdetect310.lib', which does not exist in my non-NVIDIA machine. However, I don't know how to disable this call.

If you want the MOG to work, you should

#include "opencv2/bgsegm.hpp"

and then

pMOG = bgsegm::createBackgroundSubtractorMOG();

If you want the MOG to work, you should

#include "opencv2/bgsegm.hpp"

and then

pMOG = bgsegm::createBackgroundSubtractorMOG();