(Java+processing) or (C++ + opencv)
Posted: 11 March 2013 06:55 AM   [ Ignore ]
New Member
Rank
Total Posts:  4
Joined  2013-03-11

Hi all,

First of all thank you code laboratories for making this sdk and drivers so cheap, us cannot imagine how much us have saved my life.
I just installed successfully the cl-platform sdk and the drivers for two cameras.
I need to create a lip tracker with both cameras and I seen it that the sdk works with both the Opencv or Java image libraries.

Now have a few questions that for most might sound silly, but due the limited amount of time for my project i would prefer if people from this forum with more experience would help me getting started.

- I know Java and C++ (but my memory management in C++ is not the best), which would be the best Java with processing, or C++ and openCV? is it possible at to achieve it with Java+processing?

- Does the sdk record sound or do i need an external microphone for this?

- Will this wrapper work with imageJ or is processing better for my project?

- Since i am new with image frameworks, and motion detection and is it easier to add marks (like round blue marks around the lips) or to attempt detection of the full lips?

- Is there any books/online resources that would allow me to get started with motion tracking development?

I know that asking these three questions might sound cheeky and they might even already be in this forum, however i have attempted to find my answers with no success at all, i would appreciate if it would be possible at least an answer to the first or the first two questions, since these are key to my needs.

I would appreciate any help that us could give me.

Ricardo

Profile
 
 
Posted: 13 March 2013 05:30 AM   [ Ignore ]   [ # 1 ]
New Member
Rank
Total Posts:  4
Joined  2013-03-11

I have noticed that a lot forum posts are left with no replies, so i am leaving my findings related to my questions to who may stumble upon this.

I have decided to use processing and CLEye there are wrappers and blob detection tools available online for processing and since my framework uses java processing is more handy to me than opencv (there is a wrapper for opencv but this only offers limited functionality).

I found an excellent book for processing with a motion detection chapter:
http://www.learningprocessing.com/
and this guy offer a blob detection tool for processing:
http://thomasdiewald.com/blog/?p=1107
and post with more info and code:
http://forum.processing.org/topic/new-blob-detection-and-analysis-image-labelling-and-segmentation-library-for-processing

Thanks

Profile
 
 
Posted: 13 March 2013 12:25 PM   [ Ignore ]   [ # 2 ]
Administrator
Avatar
RankRankRankRank
Total Posts:  585
Joined  2009-09-17

Hi Ricardo,

Here are the answers related to the SDK:

-The SDK will work with processing but for best performance you should use C/C++ with OpenCV.
-The SDK does not record the sound. This functionality is automatically provided by your OS.

Ricardo Duarte - 11 March 2013 06:55 AM

Hi all,

First of all thank you code laboratories for making this sdk and drivers so cheap, us cannot imagine how much us have saved my life.
I just installed successfully the cl-platform sdk and the drivers for two cameras.
I need to create a lip tracker with both cameras and I seen it that the sdk works with both the Opencv or Java image libraries.

Now have a few questions that for most might sound silly, but due the limited amount of time for my project i would prefer if people from this forum with more experience would help me getting started.

- I know Java and C++ (but my memory management in C++ is not the best), which would be the best Java with processing, or C++ and openCV? is it possible at to achieve it with Java+processing?

- Does the sdk record sound or do i need an external microphone for this?

- Will this wrapper work with imageJ or is processing better for my project?

- Since i am new with image frameworks, and motion detection and is it easier to add marks (like round blue marks around the lips) or to attempt detection of the full lips?

- Is there any books/online resources that would allow me to get started with motion tracking development?

I know that asking these three questions might sound cheeky and they might even already be in this forum, however i have attempted to find my answers with no success at all, i would appreciate if it would be possible at least an answer to the first or the first two questions, since these are key to my needs.

I would appreciate any help that us could give me.

Ricardo

Profile
 
 
Posted: 13 March 2013 01:29 PM   [ Ignore ]   [ # 3 ]
New Member
Rank
Total Posts:  4
Joined  2013-03-11

Hi,

Thank you very much for the reply, i done a test with processing and the performance seemed ok, (but i only tested the camera nothing else), I am still investigating, as the use of java is mainly because my framework uses java and i could attempt real facial capture and port the movements to the 3D model, at the same time (but if performance is not great anyways and i will eventually need to use 3 cameras to capture different angles, maybe processing wont cut it).

However if i use opencv i seen that ccv and a few other frameworks would give me a better starting point for what i need, but i couldn’t find all the dlls that come with SDK for the C++ setup.

Could you tell me how do i setup my visual studio environment to access the sdk? (or point to a webpage as i havent found nothing yet) I am asking this because the CLEyeMulticam.dll was installed at the windows folder “C:\Windows\SysWOW64\CLEyeMulticam.dll”, but it seems that there is a directshow dll file somewhere,  where i can find this file, is anymore files that i will need?Where does the installer copies these files?

Do code laboratories have some doxygen (or similar) documentation page with the installer or something, in your documentation page i couldnt see any links to it, if you could point me at all the dlls i would appreciate it, i will build documentation from there.

In the cl-eye test i done some video recording with sound (using the cl-eye mic) to test at 60fps and 120fps with just one camera and the sound was not synced at all is this a bug with Cl-eye test application or will it be hard to sync sound and video? Will sound recorded from my sound card solve these issues (I have a dedicated sound card)?

Thank you for your reply, i really appreciate your help i only have a few months to get a facial tracking application working so time is really short, which wont give me the choice to test processing and opencv for comparison, as dont have anytime to waste.

Ricardo

Profile
 
 
 
 


RSS 2.0     Atom Feed