andrei.toader's profile - activity

Did you manage to find the answer or still searching?

Hi all,

Someone happens to know a way of using BOWImgDescriptor with ORB?. Any solution is well received. I read the following link but no solution or workaround is provided. http://answers.opencv.org/question/17460/how-to-use-bag-of-words-example-with-brief/

Thanks in advance.

Thanks a lot. I understood meanwhile by taking a few examples on paper. Though, I am not sure we are talking about the same idea, because I don't mean bag of words/visual words but background subtraction using codebook.

Hi all,

Can someone explain me in detail how the codebook algorithm works?
I read about how it works but I am a bit confused about the entire process starting from how is a codeword defined and how codebooks are computed. So far I have read a part from the book Learning OpenCV(Oreilly) and from the following paper:
Real-time foreground–background segmentation using codebook model

Thanks in advance, Andrei

Though this is not a real answer I will try as best I can to give you a helpful answer.
I will try to summarize and be very explicit.
This is my current topic and it's not even close to easy. I've tried several things and concluded that it's too big to jump directly on ANDROID so I am trying to make a good hand gesture detector on PC then think later for Android.
My first(but last in terms of trying) approach for this is to create a good skin classifier (which is ongoing) and then based on the hand contour got from the skin classifier try to use convexity defects and some geometric calculations(angles, distances) to get the number of fingers shown onto the screen. Right now if the contour is right, I detect quite accurate the fingers. However, because I am using thresholds for rgb or hsv or ycrcb doesn't work well all the time so I need something way better.
A starting point for me is/was:
http://www.morethantechnical.com/2013/03/05/skin-detection-with-probability-maps-and-elliptical-boundaries-opencv-wcode/
I would suggest to try and study, if you like this approach of course, hand segmentation techniques and then once you got it working move to the next step, finding the number of fingers shown onto the screen.
My second approach(and first try) for this was(didn't work well) using a concept called Visual categorization with bag of keypoints. Starting point:
Simple Object Classifier using Bag of Words
I tried a lot to make it work but because hand is different than static objects I didn't. After investigating a lot I realized that one solution would be to use shape context instead of using points to train - but I didn't try it because of the lack of time, so I can't say for sure that it will work. You have to research that on your own. After several tests and "tunings" the best accuracy was 46% and the speed is 2FPS on PC.
Hope my answer is at least of help to you.

Good luck

I have little knowledge with Arduino (just bought one for experiments but didn't have time to work much on it), therefore I can't help you much with that. Maybe there is a problem with the serial C++ library.

I didn't try but maybe using shape context.

http://answers.opencv.org/question/1668/shape-context-implementation-in-opencv/

Your algorithm can either be SVM, Neural Network, Bayesian, etc. This thing you need yourself to analyze. Btw, about coding, don't expect someone to do the work for you. Start searching, shape detection, shape detection opencv, machine learning for shape detection(+opencv) and so on.

I got this error in opencv android also and I discovered that I must create all my instances after I receive some callback from opencv. I used the onCameraPreview(not sure if this is this the correct name). It probably has something to do with multi-threading, meaning it's possible that Opencv lib isn't initialized by the time you are making the call to create the instance (not sure if this is the case, I am just guessing). Hope it helps.

Download the opencv libraries - aka 2.4.8 in your case.

1. Set up the environment variables - properly - x86 and x64 separately. The dlls are located in the \build\x86 or x64\vcX(X can be 9, 10, etc)\bin\

2. Set the opencv libs properly in visual studio.

A good example can be found here: http://opencv-srf.blogspot.co.at/2013/05/installing-configuring-opencv-with-vs.html

For detecting features in real-time try to look at ORB or FAST. I didn't try to use them on the Android device but I made a few comparisons on PC. The results were that FAST is faster (way faster - 25ms on average (there are a few parameters you can tune)) and the accuracy is the same if not better. The bad thing with fast is that you need to use with a descriptor (if you need to) and because there was a weird bug at that time in the ORB descriptor class I used it with SURF. My suggestion would be to try to use FAST and ORB for feature detection and as for descriptor, ORB and FREAK (didn't use though, just read a bit about it).

Not sure if I am right completely but looks like this one: http://answers.opencv.org/question/27879/error-compiling-opencv-in-vs2010/

http://courses.csail.mit.edu/6.006/spring11/lectures/lec02.pdf. Hope it's relevant to what you need.

Hi all, I am working on a project using opencv4android and I notice these two options available when converting from rgb2hls. Can someone tell me the difference between the above?

thanks, Andrei

I worked a bit with this android library and the only difference I saw was in the face detector where there are these two options, native camera and java. Java detector at times seemed a bit faster. If I have to conclude, I think it's about optimization. By optimization I mean that if you decide to implement yourself it may be possible that your implementation is not as optimized as theirs.

Not sure but this may be relevant to you (Aside from configuring properly the opencv environment in vs studio). http://msdn.microsoft.com/en-us/library/ms241613(v=vs.110).aspx

didn't work?

Plus try to run opencv_create samples from command prompt

Try, restarting the vs after setting up the env var.

I use \build\x86\vc10\lib not staticlib You can see these two pictures http://imgbin.org/index.php?page=image&id=16648 and the env variables http://imgbin.org/index.php?page=image&id=16649

I hope it helps.

You got it working?

It seems for some reason that you don't have them properly setup the dlls, not sure...try to run this method from cmd and see what you get: opencv_createsamples. If you have something configured(some dlls I mean) you should see something like usage:opencv_createsamples...etc

I think you need to add the bin opencv\build\x86\vc12 to the path so that it works on visual studio 2013 too. Exactly as you did for vs 2010

I can't help you with that. I only used on 2010 and 2012. Both without problems - as I stated before. I hope you know that you have different libs for different vs versions. E.g vc10 for visual studio 10, vc11 for visual studio 12 or smth like that, sorry if I got the names wrong.

That is simple. go to Linker/General and disable Incremental Linkining: "Enable Incremental Linking: No(/INCREMENTAL:NO)".

If you read the part :

"In VS2010:

At C/C++ "Additional include directories": S:\OpenCV\build\include"

You should understand that initially I thought he is missing some configurations so yeah, I tried to give him a proper configuration to start from. I don't care about down-votes but I do care about people being fair (a lot).