cvcell-devel

Re: [Cvcell-devel] Submiting a region mask

[Hiroki S. <hi...@ea...>]

Hello, Flavio

>>>>> In <483...@gm...> 
>>>>>	Flavio Percoco Premoli <fla...@gm...> wrote:

> I'm working on a project that uses openCv-cell.

Thank you for using our product.

> In this moment I'm using the cvGoodFeaturesToTrack function to track the 
> good features, but i would like to pass a mask ( a specific region ) 
> where to track the points. The thing is that i don't really know how to 
> pass the region. I tried passing the points but it didn't work.

The cvGoodFeaturesToTrack function has the mask argument, which is a 
region of interest, described in cv.h.

CVAPI(void)  cvGoodFeaturesToTrack( const CvArr* image, CvArr* eig_image,
                                   CvArr* temp_image, CvPoint2D32f* corners,
                                   int* corner_count, double  quality_level,
                                   double  min_distance,
                                   const CvArr* mask CV_DEFAULT(NULL),
                                   int block_size CV_DEFAULT(3),
                                   int use_harris CV_DEFAULT(0),
                                   double k CV_DEFAULT(0.04) );

Do you mean that the function cannot treat the mask argument correctly?
If you mean this, this may be a bug in OpenCV.
Because the function haven't been implemented in the cvcell project yet.

> The exact thing i would like to do is get the eyebrows of a face, so i 
> guess i can use facedetect to get the face and then in that region track 
> the points (this is what I'm trying to do).  If there is a better way to 
> track specific parts of the face ( eyebrows, nose, ears, mouth ) I would 
> be really glad if you let me know this.

http://alereimondo.no-ip.org/OpenCV/34
This site may work for your purpose. This is a collection of
haar cascade classifiers that detect face, eyes, nose et al.

Sincerely,

-- 
Hiroki Sugano
E-mail: hi...@ea...
Department of Communications and Computer Engineering
Graduate School of Informatics, Kyoto University, Japan

[Cvcell-devel] Submiting a region mask

[Flavio P. P. <fla...@gm...>]

Greetings!!

I'm working on a project that uses openCv-cell.

In this moment I'm using the cvGoodFeaturesToTrack function to track the 
good features, but i would like to pass a mask ( a specific region ) 
where to track the points. The thing is that i don't really know how to 
pass the region. I tried passing the points but it didn't work.

The exact thing i would like to do is get the eyebrows of a face, so i 
guess i can use facedetect to get the face and then in that region track 
the points (this is what I'm trying to do).  If there is a better way to 
track specific parts of the face ( eyebrows, nose, ears, mouth ) I would 
be really glad if you let me know this.

I hope you can help me!!

Thanks

-- 
Flavio Percoco Premoli, A.K.A. [Flaper87]
http://www.flaper87.org
Usuario Linux registrado #436538
Geek by nature, Linux by choice, Archer of course.
Key Fingerprint: 2AAC 589E 6E61 98FC 1D80 D36B 3094 6CEA 26FD 9E19
The Solution to everything:
python -c "from struct import pack; print  pack('5b', (41*len('99')), pow(8,2)+20, 4900**0.5, range(78)[-1], 10)"

Re: [Cvcell-devel] PS3 with IP cameras

[Luciano O. <luc...@gm...>]

Hi Nakamura,

In which file of libspe source, did you see the error EINTR? I just found 
EFAULT or EPERM.

Thanks in advance
Luciano

On Thursday 24 April 2008 13:18:44 Takashi Nakamura wrote:
> Hi Luciano,
>
> > I am trying to work a Prosilicat gigabit ethernet camera with
> > opencv-cell. Now, everything works, except that when I try to acquire a
> > frame from the camera using any of the SPE, I get my system crash with a
> > error in spe_allocator.c:
>
> I tried some test with web camera, and tests work fine.
> I don't know about 'prosilica g750' and gigabit ether camera,
> so, I can't guess cause of your error.
>
> According to 'SPE_Runtime_Management_API_v2.2.pdf',
> Stopping of SPE with error code 'SPE_RUNTIME_FATAL' means
> fatal operating system errors is occured.
> And, according to sources of libspe2,
> stopinfo means 'errno' when stop with SPE_RUNTIME_FATAL.
>
> In other words, 'spe_runtime_fatal=4' means
> SPU is stopped with EINTR, but I don't know about it well.
>
>
> When I have more infomation, I may understand your problem.
> Would you check what API is broken, and please tell me more information
> about passed image, such as width, height, channel and step.
>
> Regards,



-- 
Luciano Oliveira
http://www.lucianooliveira.eti.br

Re: [Cvcell-devel] PS3 with IP cameras

[Luciano O. <luc...@gm...>]

Hi Nakamura,

One more thing: we couldn't find the _cvcell.h. Can you tell me where is it?
We are trying to integrate the camera SDK into the opencv-cell lib, is it 
possible by using just a .so file (lib, without src code, but with headers)?

Best regards
Luciano Oliveira



On Thursday 24 April 2008 13:18:44 Takashi Nakamura wrote:
> Hi Luciano,
>
> > I am trying to work a Prosilicat gigabit ethernet camera with
> > opencv-cell. Now, everything works, except that when I try to acquire a
> > frame from the camera using any of the SPE, I get my system crash with a
> > error in spe_allocator.c:
>
> I tried some test with web camera, and tests work fine.
> I don't know about 'prosilica g750' and gigabit ether camera,
> so, I can't guess cause of your error.
>
> According to 'SPE_Runtime_Management_API_v2.2.pdf',
> Stopping of SPE with error code 'SPE_RUNTIME_FATAL' means
> fatal operating system errors is occured.
> And, according to sources of libspe2,
> stopinfo means 'errno' when stop with SPE_RUNTIME_FATAL.
>
> In other words, 'spe_runtime_fatal=4' means
> SPU is stopped with EINTR, but I don't know about it well.
>
>
> When I have more infomation, I may understand your problem.
> Would you check what API is broken, and please tell me more information
> about passed image, such as width, height, channel and step.
>
> Regards,

Re: [Cvcell-devel] PS3 with IP cameras

[Luciano O. <luc...@gm...>]

Hi Nakamura,

Let me try to explain in more details:

To acquire frame from G750C gigabit ethernet camera, we must use a lib 
(PvApi.so and we don't have the source code, only the header files). What we 
are trying to do is: create a program with -lpvApi, -lcv -highgui, using also 
the patches that you created to run it in cell.

In this way, when we run this program, we get some frames (752x480) and it 
crashes with that SPE error 4. 

Inside our program, we just: 
- acquire a frame, 
- convert it to an IplImage, 
- and show in the screen (cvShowImage) - nothing more, no processing, nothing. 

Interestingly, if we just link the PvApi.so to a program without any OpenCv 
function, this error still happens (example: taking the camera program 
capture and link it to the opencv cell lib, without any opencv function)

Any idea?

Best regards
Luciano Oliveira



On Thursday 24 April 2008 13:18:44 Takashi Nakamura wrote:
> Hi Luciano,
>
> > I am trying to work a Prosilicat gigabit ethernet camera with
> > opencv-cell. Now, everything works, except that when I try to acquire a
> > frame from the camera using any of the SPE, I get my system crash with a
> > error in spe_allocator.c:
>
> I tried some test with web camera, and tests work fine.
> I don't know about 'prosilica g750' and gigabit ether camera,
> so, I can't guess cause of your error.
>
> According to 'SPE_Runtime_Management_API_v2.2.pdf',
> Stopping of SPE with error code 'SPE_RUNTIME_FATAL' means
> fatal operating system errors is occured.
> And, according to sources of libspe2,
> stopinfo means 'errno' when stop with SPE_RUNTIME_FATAL.
>
> In other words, 'spe_runtime_fatal=4' means
> SPU is stopped with EINTR, but I don't know about it well.
>
>
> When I have more infomation, I may understand your problem.
> Would you check what API is broken, and please tell me more information
> about passed image, such as width, height, channel and step.
>
> Regards,

Re: [Cvcell-devel] PS3 with IP cameras

[Takashi N. <nak...@fi...>]

Hi Luciano,

> I am trying to work a Prosilicat gigabit ethernet camera with opencv-cell. 
> Now, everything works, except that when I try to acquire a frame from the 
> camera using any of the SPE, I get my system crash with a error in 
> spe_allocator.c:

I tried some test with web camera, and tests work fine.
I don't know about 'prosilica g750' and gigabit ether camera,
so, I can't guess cause of your error.

According to 'SPE_Runtime_Management_API_v2.2.pdf', 
Stopping of SPE with error code 'SPE_RUNTIME_FATAL' means
fatal operating system errors is occured.
And, according to sources of libspe2, 
stopinfo means 'errno' when stop with SPE_RUNTIME_FATAL.

In other words, 'spe_runtime_fatal=4' means 
SPU is stopped with EINTR, but I don't know about it well.


When I have more infomation, I may understand your problem.
Would you check what API is broken, and please tell me more information 
about passed image, such as width, height, channel and step.

Regards,
-- 
Takashi<nak...@fi...>

[Cvcell-devel] PS3 with IP cameras

[Luciano O. <luc...@gm...>]

Hi,

I am trying to work a Prosilicat gigabit ethernet camera with opencv-cell. 
Now, everything works, except that when I try to acquire a frame from the 
camera using any of the SPE, I get my system crash with a error in 
spe_allocator.c:

SPE_RUNTIME_FATAL = 4

In my program, I am using part of the acquisition module provided by Prosilica 
and copying the frame into IplImage struct and then calling cvShow to view 
the frames from the camera. With CVCELL_SPENUM=0, everything works fine, but 
the goal is to use SPE's.

Any help?

Best regards
-- 
Luciano Oliveira
http://www.lucianooliveira.eti.br

Re: [Cvcell-devel] cvInRangeS

[Takashi N. <nak...@fi...>]

Hi Joris,

> I'm trying to get the camshiftdemo optimized for the PS3.
> I found out cvInRangeS takes a lot of time. Cvcell has optimized this
> function, but cvcell is not using the optimized cvInRangeS, because of this:
> in cxcmp.cpp (ppu source folder, line 379):
>     if( ( spenum > 0 ) && CHECK_ARY_SPU( src1 ) &&
>     CHECK_ARY_SPU( src2 ) && CHECK_ARY_SPU( src3 ) && CHECK_ARY_SPU( dst )
> &&
>      (CV_MAT_CN(src1->type) == 1) )
>     {
>         cvInRange_spu(src1, src2, src3, dst);
>         return;
>     }
> 
> I used the debugger and found out that CV_MAT_CN(src1->type)==3, so it will
> not use the spu's.
> Is there someone that could tell me what that is and how I could get this
> function to work on de spu's?

cvInRange on SPU for 3 channel matrix is not yet implemented.
cvInRange for 3 channel matrix and cvInRange for
a single channel matrix are differed.
Result of cvInRange for 3 channel matrix is one that is bit AND-ed 
result of cvInRange of each channel.
So, current implementation of cvInRange can't
apply to 3 channel matrix.

If you need that API, you need implement that.
(Sorry, I want to implement it, but I'm busy...)



Implementing cvInRange on SPU for three channels will be
achieved by follow instructions:

 - Split your matrix using cvSplit. 
 - Apply cvInRangeS to each channel of the matrix.
 - Apply cvAnd to result of cvInRangeS.

The follow is the pseudocode of outline of that instructions:
========================================
CvMat tmp_b;
CvMat tmp_c[3];

cvSplit(src1, tmp_c[0], tmp_c[1], tmp_c[2]);

cvInRange(tmp_c[0], lower, upper, dst);
cvInRange(tmp_c[1], lower, upper, tmp_b);
cvAnd(dst, tmp_b, dst);
cvInRange(tmp_c[2], lower, upper, tmp_b);
cvAnd(dst, tmp_b, dst);
========================================

But, this implementation is not efficient, because 
that instructions need overhead of invoking cvAnd and cvSplit.

If you need more performance of cvInRange, 
try implementation cvInRange_spu for three channel matrix.

Thanks,
Takashi
-- 
Takashi Nakamura<nak...@fi...>

[Cvcell-devel] cvInRangeS

[Joris D'H. <jor...@gm...>]

Hi,

I'm trying to get the camshiftdemo optimized for the PS3.
I found out cvInRangeS takes a lot of time. Cvcell has optimized this
function, but cvcell is not using the optimized cvInRangeS, because of this:
in cxcmp.cpp (ppu source folder, line 379):
    if( ( spenum > 0 ) && CHECK_ARY_SPU( src1 ) &&
    CHECK_ARY_SPU( src2 ) && CHECK_ARY_SPU( src3 ) && CHECK_ARY_SPU( dst )
&&
     (CV_MAT_CN(src1->type) == 1) )
    {
        cvInRange_spu(src1, src2, src3, dst);
        return;
    }

I used the debugger and found out that CV_MAT_CN(src1->type)==3, so it will
not use the spu's.
Is there someone that could tell me what that is and how I could get this
function to work on de spu's?

Thanks,

Joris

[Cvcell-devel] cvInRangeS

[Joris D'H. <jor...@gm...>]

Hi,

I'm trying to get the camshiftdemo optimized for the PS3.
I found out cvInRangeS takes a lot of time. Cvcell has optimized this
function, but cvcell is not using the optimized cvInRangeS, because of this:
in cxcmp.cpp (ppu source folder, line 379):
    if( ( spenum > 0 ) && CHECK_ARY_SPU( src1 ) &&
    CHECK_ARY_SPU( src2 ) && CHECK_ARY_SPU( src3 ) && CHECK_ARY_SPU( dst )
&&
    (CV_MAT_CN(src1->type) == 1) )
    {
        cvInRange_spu(src1, src2, src3, dst);
        return;
    }

I used the debugger and found out that CV_MAT_CN(src1->type)==3, so it will
not use the spu's.
Is there someone that could tell me what that is and how I could get this
function to work on de spu's?

Thanks,

Joris

[Cvcell-devel] cvcell mailing list test

[Takashi N. <nak...@fi...>]

This is test mail.

-- 
nakamura