Ask Your Question

the___doctor's profile - activity

2014-07-03 21:48:30 -0600 commented answer How to specify mask for grabCut in C++

Very helpful! I have not seen a snippet online that did the grabcut as well! Good job!

2014-07-03 21:48:02 -0600 received badge  Supporter (source)
2014-06-21 01:45:04 -0600 asked a question Average Nearby Detected pixels into single point

Is there a way to average nearby detects pixels (extracted using canny and findcontours). The picture below shows the points of the hand, and i am trying to find a way to average the detected averages in each location to a sing Point.

example image here!

Any help would be much appreciated, Thanks in advance!

2014-06-16 22:20:02 -0600 asked a question Canny - find absence of contours, find gap

Hi, i am trying to find a way to find an absence of contours in in vector<vector<point>>.

(a pictorial example of the value i am trying to find). The region (average) with the least values/contours is what i am trying to retrieve (where - is a value/contour).

---------- -   -   -             ---    ----
---- - -----------------          ----------
------------- -------- --        - ---------

I apologize for the poorly worded question, if any clarification is needed, please comment.

2014-06-14 02:10:24 -0600 commented question Order of image manipulation

if you post comment as answer, i will accept, cheers.

2014-06-14 01:42:49 -0600 received badge  Editor (source)
2014-06-14 01:42:28 -0600 asked a question Order of image manipulation

I am trying to write a program that will give an outline of an set object in an image like so:

cvtColor( frameToUse, frameToUse, CV_BGR2GRAY );          //1
equalizeHist( frameToUse, frameToUse );                   //2
blur(frameToUse, frameToUse, Size(9,9), Point(-1,1), 4);  //3
Canny( frameToUse, frameToUse, 40, 3*35, 3 );             //4

I am wondering, what is the correct order to put these functions in to produce the results most efficiently. (ill closely watch comments if clarification is needed)

2014-06-12 15:33:37 -0600 answered a question The Active Appearance Model for iOS

if you want it like the youtube video you linked, it could be done quite simply, all would have to to do (in opencv) is detect the face using one the the face haarcascades like profileface_alt2_haarcascade.xml and (if you want crop the region of the face), and inside the ROI, check for smiles using the smile_haarcascade.xml.

Using this information you could easily identify on the screen if a person is smiling or not

2014-06-06 01:37:02 -0600 commented question train memory error

did you find the answer to this question? I am having the exact same issue.