Ask Your Question

beahacker's profile - activity

2016-06-05 13:16:10 -0500 received badge  Necromancer (source)
2015-09-15 06:17:26 -0500 commented answer TLD Tracker (aka Predator)

Hi Mubb, yes it is in OpenCV 3.0

2015-07-22 08:17:38 -0500 received badge  Necromancer (source)
2015-06-06 03:21:25 -0500 commented answer Cannot generate matlab bindings in OpenCV

@berak: Hi, I did what you said, it passed Cmake step, but cannot build. The error log: [ 13%] Generating Matlab source files [ 13%] Built target opencv_test_ximgproc_pch_dephelp [ 13%] Built target pch_Generate_opencv_hal usage: gen_matlab.py [-h] [--jinja2 JINJA2] [--hdrparser HDRPARSER] [--moduleroot [MODULEROOT [MODULEROOT ...]]] [--modules [MODULES [MODULES ...]]] [--extra [EXTRA [EXTRA ...]]] [--outdir OUTDIR] gen_matlab.py: error: unrecognized arguments: --rstparser /home/beahacker/Downloads/opencv-new/opencv/modules/java/generator make[2]: * [modules/matlab/generate.proxy] Error 2 make[1]: [modules/matlab/CMakeFiles/opencv_matlab_sources.dir/all] Error 2 make[1]: ** Waiting for unfinished jobs....

2015-06-02 23:14:26 -0500 received badge  Editor (source)
2015-06-02 23:14:02 -0500 answered a question Cannot generate matlab bindings in OpenCV

@krammer: You can download rst_parser.py here https://github.com/Itseez/opencv/blob.... I downloaded and put it in my local corresponding folder, according to this solution [http://code.opencv.org/issues/4175]. And it works for me.

2015-06-02 22:30:41 -0500 commented question Cannot generate matlab bindings in OpenCV

@StevenPuttemans: Hi, I also use branch 3.0.0-dev, if it contains a lot of bugs, but bindings should not be a problem. So sad, I cannot build it also.

2015-05-26 01:32:51 -0500 answered a question TLD Tracker (aka Predator)

Hi, In a demo https://github.com/Itseez/opencv_cont... of opencv Tracking algorithm, you can use several algorithms such as BOOSTING, TLD, MIL, MEDIANFLOW.
I tested it and found that BOOSTING algorithm is more robust to fast motion than other trackers. My dataset is sports dataset, so it contains many videos with fast motions. Other than that, STRUCK tracking is one of the best in current tracking benchmark https://sites.google.com/site/tracker.... The STRUCK implementation can be found on Github https://github.com/gnebehay/STRUCK or my adapted and modified version https://github.com/howtobeahacker/myS....

2015-05-15 00:19:38 -0500 received badge  Enthusiast
2015-05-14 00:55:06 -0500 commented answer TLD Tracker (aka Predator)

I tested MIL, BOOSTING, MEDIANFLOW, TLD. None of them is good for sports datasets.

2015-05-09 06:28:36 -0500 received badge  Teacher (source)
2015-05-08 09:39:45 -0500 commented answer Bag of Features Descriptor

A good answer!

2015-05-08 09:36:42 -0500 received badge  Supporter (source)
2015-05-08 09:26:37 -0500 answered a question How does Lukas kanade optical flow and goodFeaturesToTrack works?

Basically, GoodFeatureToTrack function computes Harris corner scores at every pixel in the image. Then it thresholds these scores to produce good features to track. What is Harris corner score, pls look at this http://en.wikipedia.org/wiki/Corner_d.... LK optical flow is quite long story to explain. You can look up it on Internet.

2015-05-07 03:14:23 -0500 answered a question What is the most effective way to access cv::Mat elements in a loop?

According to the official documentations, they suggest that the most efficient way is to get the pointer to the row first, and then just use the plain C operator []. It also saves a multiplication for each iteration.

// compute sum of positive matrix elements
// (assuming that M isa double-precision matrix)
double sum=0;
for(int i = 0; i < M.rows; i++)
{
    const double* Mi = M.ptr<double>(i);
    for(int j = 0; j < M.cols; j++)
        sum += std::max(Mi[j], 0.);
}