Skip to search.

Breaking News Visit Yahoo! News for the latest.

×Close this window

OpenCV · Open Source Computer Vision Library
  • Members Only
  • Post
  • Files
  • Photos
  • Polls
  • Members
  • Promote

The Yahoo! Groups Product Blog

Check it out!

Group Information

  • Members: 48147
  • Category: Open Source
  • Founded: Jun 20, 2000
  • Language: English
? Already a member? Sign in to Yahoo!

Yahoo! Groups Tips

Did you know...
Message search is now enhanced, find messages faster. Take it for a spin.

Messages
Message #
Search:
Advanced
 Messages Help
Messages
Topics
Hot Topics
Messages 87601 - 87631 of 90774   Oldest  |  < Older  |  Newer >  |  Newest
Messages: Show Message Summaries Sort by Date ^  
#87601 From: "munir" <munircs@...>
Date: Thu May 3, 2012 9:20 pm
Subject: Re: Reading / Writing MS Excel files in Opencv 		munircs
Send Email Send Email
  
Thanks a lot Kishor Durve.
yep it helps.........

--- In OpenCV@yahoogroups.com, Kishor Durve <kishor_durve@...> wrote:
>
> �I think you will have to create a .csv (comma separated) file and use a
conversion utility like Dataload to convert this file to .xls -�at least
this��is what I do.
>
> Kishor Durve
> 'A single dream is more powerful than a thousand realities'
>

Reply | Messages in this Topic (8)
#87602 From: "ehsan" <ehsanrza@...>
Date: Fri May 4, 2012 1:59 am
Subject: Cmake for windows 7 64 bit 		ehsanrza
Send Email Send Email
  
Hello everyone,

As many other people pointed out, it seems that it is not possible to build a
project using visual C++ 2010 64 bit and OpenCV 2.3.1
In addition, there is no Cmake for Windows 64 bits.

Any advices is appreciated!

Reply | Messages in this Topic (1)
#87603 From: "anatoly_2011" <anatolyb2011@...>
Date: Fri May 4, 2012 7:08 am
Subject: Re: Issue with GPU and GeForce GTX 680 		anatoly_2011
Send Email Send Email
  
BTW, if you add '30' to CUDA_ARCH_BIN varibable in OPenCV CMake, the code is
supposed to work faster.

BTW, I haven't had any chance to test OpenCV with Kepler.

--- In OpenCV@yahoogroups.com, Walter Lucetti <w.lucetti@...> wrote:
>
> It was my hurry... after waiting a few seconds the test application started
> and finally I got it working.
>
> Thank you
> Walter
>
> 2012/5/3 Thomas Bastiani <thom@...>
>
> > **
> >
> >
> > Hi Walter,
> >
> > Can you give us the exact revision of your CUDA Driver, toolkit and your
> > CUDA SDK?
> > Also, can you try running the deviceQuery CUDA sample to see if it
> > completes?
> >
> > Eventually, if deviceQuery freezes up, build it in debug mode and run it
> > under gdb so that we know where it's blocked.
> >
> > Cheers,
> > --
> > Thomas
> >
> >
> > On 3 May 2012 09:46, Walter Lucetti <w.lucetti@...> wrote:
> >
> >>
> >>
> >> Hi Anatoly,
> >>
> >> I updated to latest trunk (rev. 8295) and now Kepler device info is
> >> printed correctly.
> >>
> >> I tried again "performance_gpu" example and it freezes at this point:
> >>
> >> ======
> >> Device 0:  "GeForce GTX 680"  2048Mb, sm_30, 1536 cores, Driver/Runtime
> >> ver.4.20/4.20
> >>
> >>     CPU, ms   GPU, ms   SPEEDUP       DESCRIPTION
> >>
> >> matchTemplate:
> >> ======
> >>
> >> Each gpu test application seems to freeze at the beginning of computation.
> >>
> >> Walter
> >>
> >>
> >> 2012/5/3 Walter Lucetti <w.lucetti@...>
> >>
> >>> Thank you Anatoly. Tomorrow I will take the latest trunk instead of the
> >>> stable.
> >>>
> >>> This morning (Italy time) it didn't compile and I went straight to
> >>> official version.
> >>>
> >>>
> >>> 2012/5/2 anatoly_2011 <anatolyb2011@...>
> >>>
> >>>> **
> >>>>
> >>>>
> >>>> This is fixed in trunk version in r8293
> >>>>
> >>>>
> >>>> --- In OpenCV@yahoogroups.com, Walter Lucetti <w.lucetti@> wrote:
> >>>> >
> >>>> > Hi all,
> >>>> >
> >>>> > I installed new OpenCV 2.4 and compiled with latest CUDA driver,
> >>>> toolkit
> >>>> > and SDK under Linux 11.10.
> >>>> >
> >>>> > I compiled everything without any problems, but I cannot run examples
> >>>> and
> >>>> > test (they all freeze).
> >>>> >
> >>>> > The only example that gives a feedback is "performance_gpu" and I get
> >>>> this
> >>>> > error:
> >>>> >
> >>>> > "Device 0: "GeForce GTX 680" 2048MbMapSMtoCores undefined SMversion
> >>>> 3.0!"
> >>>> >
> >>>> > Any idea?
> >>>> >
> >>>> > Thank you
> >>>> > Walter
> >>>> >
> >>>>
> >>>>
> >>>
> >>
> >>
> >>
> >
> >
>

Reply | Messages in this Topic (9)
#87604 From: Walter Lucetti <w.lucetti@...>
Date: Fri May 4, 2012 7:45 am
Subject: Re: Re: Issue with GPU and GeForce GTX 680 		myzhar77
Send Email Send Email
 
Hi Anatoly,

I finally reached a good configuration that let me test the card.

The final result for "performance_gpu" is "average GPU speedup: x41.827"

I also tested "stereo_match_gpu" with "aloe" and "tsucuba"

These are the results:

ALOE about 90 FPS
--- Parameters ---
image_size: (1282, 1110)
image_channels: 1
method: BM
ndisp: 64
win_size: 19
prefilter_sobel: 0

TSUCUBA about 740 FPS (!!!)
--- Parameters ---
image_size: (384, 288)
image_channels: 1
method: BM
ndisp: 64
win_size: 19
prefilter_sobel: 0





Cheers
Walter

PS Are you the same Anatoly I am talking with in PCL mailing list?

2012/5/4 anatoly_2011 <anatolyb2011@...>
 

BTW, if you add '30' to CUDA_ARCH_BIN varibable in OPenCV CMake, the code is supposed to work faster.

BTW, I haven't had any chance to test OpenCV with Kepler.



--- In OpenCV@yahoogroups.com, Walter Lucetti <w.lucetti@...> wrote:
>
> It was my hurry... after waiting a few seconds the test application started
> and finally I got it working.
>
> Thank you
> Walter
>
> 2012/5/3 Thomas Bastiani <thom@...>
>
> > **

> >
> >
> > Hi Walter,
> >
> > Can you give us the exact revision of your CUDA Driver, toolkit and your
> > CUDA SDK?
> > Also, can you try running the deviceQuery CUDA sample to see if it
> > completes?
> >
> > Eventually, if deviceQuery freezes up, build it in debug mode and run it
> > under gdb so that we know where it's blocked.
> >
> > Cheers,
> > --
> > Thomas
> >
> >
> > On 3 May 2012 09:46, Walter Lucetti <w.lucetti@...> wrote:
> >
> >>
> >>
> >> Hi Anatoly,
> >>
> >> I updated to latest trunk (rev. 8295) and now Kepler device info is
> >> printed correctly.
> >>
> >> I tried again "performance_gpu" example and it freezes at this point:
> >>
> >> ======
> >> Device 0: "GeForce GTX 680" 2048Mb, sm_30, 1536 cores, Driver/Runtime
> >> ver.4.20/4.20
> >>
> >> CPU, ms GPU, ms SPEEDUP DESCRIPTION
> >>
> >> matchTemplate:
> >> ======
> >>
> >> Each gpu test application seems to freeze at the beginning of computation.
> >>
> >> Walter
> >>
> >>
> >> 2012/5/3 Walter Lucetti <w.lucetti@...>

> >>
> >>> Thank you Anatoly. Tomorrow I will take the latest trunk instead of the
> >>> stable.
> >>>
> >>> This morning (Italy time) it didn't compile and I went straight to
> >>> official version.
> >>>
> >>>
> >>> 2012/5/2 anatoly_2011 <anatolyb2011@...>
> >>>
> >>>> **

> >>>>
> >>>>
> >>>> This is fixed in trunk version in r8293
> >>>>
> >>>>
> >>>> --- In OpenCV@yahoogroups.com, Walter Lucetti <w.lucetti@> wrote:
> >>>> >
> >>>> > Hi all,
> >>>> >
> >>>> > I installed new OpenCV 2.4 and compiled with latest CUDA driver,
> >>>> toolkit
> >>>> > and SDK under Linux 11.10.
> >>>> >
> >>>> > I compiled everything without any problems, but I cannot run examples
> >>>> and
> >>>> > test (they all freeze).
> >>>> >
> >>>> > The only example that gives a feedback is "performance_gpu" and I get
> >>>> this
> >>>> > error:
> >>>> >
> >>>> > "Device 0: "GeForce GTX 680" 2048MbMapSMtoCores undefined SMversion
> >>>> 3.0!"
> >>>> >
> >>>> > Any idea?
> >>>> >
> >>>> > Thank you
> >>>> > Walter
> >>>> >
> >>>>
> >>>>
> >>>
> >>
> >>
> >>
> >
> >
>



Reply | Messages in this Topic (9)
#87605 From: Hemant Singh <unizen01@...>
Date: Thu May 3, 2012 12:58 pm
Subject: Re: Human Counter System HELP 		maddyhemu
Send Email Send Email
 
Try blob.. it becomes easier

btw, the image you provided have some lightning issues coz when the person comes in the scene, the light reflected (as seen in the brighter part of the image in the top is detected as same as background) so in differencing it doesnt counts. send in some more pics of the bg without person and with person if possible.

Hope this helps.

Cheers,
Hemant

On Thu, May 3, 2012 at 4:29 PM, gkhanacer <gkhanacer@...> wrote:
 

Hi, I'm Gökhan.I'm new to OpenCV.
I have a project about human counter system with opencv.
I try to remove background and extract foreground, and
I appied frame differencing.
But Result Image has a lot of noise.
I applied cvErode,cvDilate,cvSmooth and Countor Analysis.
But people's bodies aren't complete.
http://img3.imageshack.us/img3/5627/asasjt.jpg
How to assemble ? or How to take the whole image of the human body?
Should I use cvBlobLib?




--

Reply | Messages in this Topic (4)
#87606 From: Hemant Singh <unizen01@...>
Date: Thu May 3, 2012 1:00 pm
Subject: Re: HAAR CLASSIFIER BASED HAND DETECTION AND CAMSHIFT TRACKING 		maddyhemu
Send Email Send Email
 
Google it man!!

On Thu, May 3, 2012 at 11:40 AM, sawan <arsahil_2007@...> wrote:
 

HI ALL
can anybody share with me code for haar classifier based hand detection and tracking hand using camshift for gesture recognition

thankx




--


-Hemant Kumar Singh


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

“Your time is limited, so don't waste it living someone else's life. Don't be trapped by dogma - which is living with the results of other people's thinking. Don't let the noise of other's opinions drown out your own inner voice. And most important, have the courage to follow your heart and intuition. They somehow already know what you truly want to become. Everything else is secondary.”

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


Reply | Messages in this Topic (2)
#87607 From: Hemant Singh <unizen01@...>
Date: Thu May 3, 2012 1:04 pm
Subject: Re: linking game and opencv 		maddyhemu
Send Email Send Email
 
why open source... u can do it for standard window games like NFS too !!

Just associate your gesture with window keys, and there you go!!

Have Fun :)

-Hemant

On Wed, May 2, 2012 at 10:42 PM, sohaib abid <sohaib.abid@...> wrote:
 

0N APRIL 29 2012 I WROTE


Hi, i have made an opencv (c++) code for simple gesture recognition of left and right movement of hand,
Now i want to use this program to control simple left/right car movements in any open source game

Can any one of you tell me any such game...? ? ?

I have googled some games but it would be easy for me, if anyone of you can refer me a game you people have already tried to manipulate

                        thanking in advance....




I HAVE GOT A SUGGESTION TO USE TUX RACER AS A GAME, BUT NOW I WANT TO ASK YOU ANY METHOD OR LINK YOU KNOW TO PROGRAM THIS GAME TO WORK WITH MY OPENCV CODE...

               




--


-Hemant Kumar Singh


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

“Your time is limited, so don't waste it living someone else's life. Don't be trapped by dogma - which is living with the results of other people's thinking. Don't let the noise of other's opinions drown out your own inner voice. And most important, have the courage to follow your heart and intuition. They somehow already know what you truly want to become. Everything else is secondary.”

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


Reply | Messages in this Topic (2)
#87608 From: "mnowak022" <mnowak022@...>
Date: Thu May 3, 2012 1:33 pm
Subject: Re: thinning function equivelent to matlab 		mnowak022
Send Email Send Email
  
Could you send me this code?

Reply | Messages in this Topic (11)
#87609 From: "s13g3l" <s13g3l@...>
Date: Thu May 3, 2012 1:07 pm
Subject: Re: FLANN: int data instead of float data 		s13g3l
Send Email Send Email
  
In the FLANN 1.7.1 documentation for the C API it says in section 3.2:
"[...] Because there is not template support in C, there are bindings provided
for the following data types: unsigned char, int, float and double. [...]"

This means that the FLANN index supports these data types, but OpenCV 2.3.0
simply does provide the interface to all of them (only to float data).

Are there any plans for adapting the OpenCV FLANN interface to offer all data
types supported by the FLANN index?

Reply | Messages in this Topic (3)
#87610 From: Bala subramaniyan Murugesan <cmbaala@...>
Date: Thu May 3, 2012 1:35 pm
Subject: For Help in too few arguments to function ‘cvHaarDetectObjects’ 		cmbaala
Send Email Send Email
 
Hi,
  i am using OpenCV 2.4.0 when import a sample face detection program for image it shows error in function like " too few arguments to function ‘cvHaarDetectObjects’ ". If i open that method declaration in objdetect.hpp means that method have 8 arguments but my sample program have only 7 arguments. Please help me to troubleshoot this problem.



Regards,
Balasubramaniyan Murugesan.

Reply | Messages in this Topic (3)
#87611 From: "tagyro" <tagyro@...>
Date: Thu May 3, 2012 3:05 pm
Subject: Newb question 		tagyro
Send Email Send Email
  
Hi everyone!

This is my first experience with OpenCV so please be gentle :)

I need to write a function that returns the percent of similarity between two
images (loaded from the disk). The images are of different resolution and size
(one is taken with a 1.3 MP camera, the other with a 5MP camera).
I googled a bit about this but haven't found a clear answer so I thought asking
here might save some time.
Basically I need a start: what to look at, what to read, how would I go about to
write this.
I'm pretty familiar with C/C++ and JavaScript and have some basic knowledge of
Objective-C.

Any help is greatly appreciated.

Thank you,
tagyro

Reply | Messages in this Topic (6)
#87612 From: "Septioadi Anggara" <holmes_tyo_sensei@...>
Date: Thu May 3, 2012 4:13 pm
Subject: tweaking camshift.cpp source... 		holmes_tyo_s...
Send Email Send Email
  
Hi everyone,

I'm sorry if my english are bad because I'm not a native..


my problem is that I still don't understand what's this line of code means ..

cur_rect = cv::Rect(cur_rect) & cv::Rect(0, 0, mat->cols, mat->rows);
if( cv::Rect(cur_rect) == cv::Rect() )
         {
             cur_rect.x = mat->cols/2;
             cur_rect.y = mat->rows/2;
         }

it's in the camshift.cpp source..


thank you

Reply | Messages in this Topic (1)
#87613 From: "trupswithlove" <truptivk@...>
Date: Thu May 3, 2012 5:14 pm
Subject: OpenCV2.4.0 - facerec_demo runtime error 		trupswithlove
Send Email Send Email
 

Hello,

I am new to OpenCV. I am specifically exploring Face recognition capabilities added in OpenCV2.4beta version. After compiling and executing samples/cpp/facerec_demo.cpp, I get following runtime error.

OpenCV Error: Assertion failed (depth == CV_8U || depth == CV_16U || depth == CV_32F) in cvtColor, file /home/truptivk/Desktop/Face Recognition /OpenCV-2.4.0-beta2/modules/imgproc/src/color.cpp, line 3143
terminate called after throwing an instance of 'cv::Exception'
what():  /home/truptivk/Desktop/Face Recognition /OpenCV-2.4.0-beta2/modules/imgproc/src/color.cpp:3143: error: (-215) depth == CV_8U || depth == CV_16U || depth == CV_32F in function cvtColor
Aborted (core dumped)

After dumping the depth value it shows CV_64F.

I have created csvfile.ext for AT&T database. (Images have 8-bit grey levels.) and I am using Ubuntu 12.04 32 bit version. I am clueless from where its picking up CV_64F.

Please help me to rectify this error. What needs to be one in order to avoid this?

Thanks & Regards,

Trupti Kulkarni



Reply | Messages in this Topic (3)
#87614 From: trupti kulkarni <truptivk@...>
Date: Thu May 3, 2012 6:03 pm
Subject: OpenCV2.4.0 - facerec_demo runtime error 		trupswithlove
Send Email Send Email
 

Hello, 

I am new to OpenCV. I am specifically exploring Face recognition capabilities added in OpenCV2.4 version. After compiling and executing samples/cpp/facerec_demo.cpp, I get following runtime error.

OpenCV Error: Assertion failed (depth == CV_8U || depth == CV_16U || depth == CV_32F) in cvtColor, file /home/truptivk/Desktop/Face Recognition /OpenCV-2.4.0-beta2/modules/imgproc/src/color.cpp, line 3143
terminate called after throwing an instance of 'cv::Exception'
what(): /home/truptivk/Desktop/Face Recognition /OpenCV-2.4.0-beta2/modules/imgproc/src/color.cpp:3143: error: (-215) depth == CV_8U || depth == CV_16U || depth == CV_32F in function cvtColor
Aborted (core dumped)

After dumping the depth value it shows CV_64F.

I have created csvfile.ext for AT&T database. (Images have 8-bit grey levels.) and I am using Ubuntu 12.04 32 bit version. I am clueless from where its picking up CV_64F.

Please help me to rectify this error. What needs to be one in order to avoid this? 

Thanks & Regards, 

Trupti Kulkarni


Reply | Messages in this Topic (3)
#87615 From: "mcleod_sterling" <mcleod_sterling@...>
Date: Thu May 3, 2012 8:37 pm
Subject: Installing in Debian Squeeze 		mcleod_sterling
Send Email Send Email
  
I am trying to install OpenCV in Debian Squeeze. However, whenever I try to
build the library, it fails with many errors in the zconf.h file. A total list
is -

[code]
In file included from /usr/include/libpng12/png.h:514,
                  from /home/sterling/OpenCV-2.1.0/src/highgui/grfmt_png.cpp:57:
/usr/include/zlib.h:80: error: typedef `voidpf' is initialized (use decltype
instead)
/usr/include/zlib.h:80: error: `alloc_func' was not declared in this scope
/usr/include/zlib.h:81: error: typedef `free_func' is initialized (use decltype
instead)
/usr/include/zlib.h:81: error: `voidpf' was not declared in this scope
/usr/include/zlib.h:81: error: `voidpf' was not declared in this scope
/usr/include/zlib.h:86: error: expected `;' before `*' token
/usr/include/zlib.h:87: error: `uInt' does not name a type
/usr/include/zlib.h:88: error: `uLong' does not name a type
/usr/include/zlib.h:90: error: expected `;' before `*' token
/usr/include/zlib.h:91: error: `uInt' does not name a type
/usr/include/zlib.h:92: error: `uLong' does not name a type
/usr/include/zlib.h:97: error: `alloc_func' does not name a type
/usr/include/zlib.h:98: error: `free_func' does not name a type
/usr/include/zlib.h:99: error: `voidpf' does not name a type
/usr/include/zlib.h:102: error: `uLong' does not name a type
/usr/include/zlib.h:103: error: `uLong' does not name a type
/usr/include/zlib.h:114: error: `uLong' does not name a type
/usr/include/zlib.h:117: error: expected `;' before `*' token
/usr/include/zlib.h:118: error: `uInt' does not name a type
/usr/include/zlib.h:119: error: `uInt' does not name a type
/usr/include/zlib.h:120: error: expected `;' before `*' token
/usr/include/zlib.h:121: error: `uInt' does not name a type
/usr/include/zlib.h:122: error: expected `;' before `*' token
/usr/include/zlib.h:123: error: `uInt' does not name a type
/usr/include/zlib.h:578: error: expected `,' or `...' before `*' token
/usr/include/zlib.h:684: error: `uLong' does not name a type
/usr/include/zlib.h:784: error: expected `,' or `...' before `*' token
/usr/include/zlib.h:1050: error: `uLong' does not name a type
/usr/include/zlib.h:1102: error: `Bytef' was not declared in this scope
/usr/include/zlib.h:1102: error: `dest' was not declared in this scope
/usr/include/zlib.h:1102: error: `uLongf' was not declared in this scope
/usr/include/zlib.h:1102: error: `destLen' was not declared in this scope
/usr/include/zlib.h:1102: error: expected primary-expression before `const'
/usr/include/zlib.h:1102: error: `uLong' was not declared in this scope
/usr/include/zlib.h:1102: error: initializer expression list treated as compound
expression
/usr/include/zlib.h:1116: error: `Bytef' was not declared in this scope
/usr/include/zlib.h:1116: error: `dest' was not declared in this scope
/usr/include/zlib.h:1116: error: `uLongf' was not declared in this scope
/usr/include/zlib.h:1116: error: `destLen' was not declared in this scope
/usr/include/zlib.h:1116: error: expected primary-expression before `const'
/usr/include/zlib.h:1116: error: `uLong' was not declared in this scope
/usr/include/zlib.h:1116: error: expected primary-expression before `int'
/usr/include/zlib.h:1116: error: initializer expression list treated as compound
expression
/usr/include/zlib.h:1132: error: `uLong' does not name a type
/usr/include/zlib.h:1139: error: `Bytef' was not declared in this scope
/usr/include/zlib.h:1139: error: `dest' was not declared in this scope
/usr/include/zlib.h:1139: error: `uLongf' was not declared in this scope
/usr/include/zlib.h:1139: error: `destLen' was not declared in this scope
/usr/include/zlib.h:1139: error: expected primary-expression before `const'
/usr/include/zlib.h:1139: error: `uLong' was not declared in this scope
/usr/include/zlib.h:1139: error: initializer expression list treated as compound
expression
/usr/include/zlib.h:1165: error: `voidp' does not name a type
/usr/include/zlib.h:1190: error: `gzFile' does not name a type
/usr/include/zlib.h:1209: error: `gzFile' was not declared in this scope
/usr/include/zlib.h:1209: error: expected primary-expression before `unsigned'
/usr/include/zlib.h:1209: error: initializer expression list treated as compound
expression
/usr/include/zlib.h:1226: error: `gzFile' was not declared in this scope
/usr/include/zlib.h:1226: error: expected primary-expression before `int'
/usr/include/zlib.h:1226: error: expected primary-expression before `int'
/usr/include/zlib.h:1226: error: initializer expression list treated as compound
expression
/usr/include/zlib.h:1235: error: `gzFile' was not declared in this scope
/usr/include/zlib.h:1235: error: `voidp' was not declared in this scope
/usr/include/zlib.h:1235: error: expected primary-expression before `unsigned'
/usr/include/zlib.h:1235: error: initializer expression list treated as compound
expression
/usr/include/zlib.h:1251: error: `gzFile' was not declared in this scope
/usr/include/zlib.h:1251: error: `voidpc' was not declared in this scope
/usr/include/zlib.h:1251: error: expected primary-expression before `unsigned'
/usr/include/zlib.h:1251: error: initializer expression list treated as compound
expression
/usr/include/zlib.h:1259: error: `gzFile' was not declared in this scope
/usr/include/zlib.h:1259: error: expected primary-expression before `const'
/usr/include/zlib.h:1259: error: expected primary-expression before `...' token
/usr/include/zlib.h:1259: error: initializer expression list treated as compound
expression
/usr/include/zlib.h:1274: error: `gzFile' was not declared in this scope
/usr/include/zlib.h:1274: error: expected primary-expression before `const'
/usr/include/zlib.h:1274: error: initializer expression list treated as compound
expression
/usr/include/zlib.h:1282: error: `gzFile' was not declared in this scope
/usr/include/zlib.h:1282: error: expected primary-expression before `char'
/usr/include/zlib.h:1282: error: expected primary-expression before `int'
/usr/include/zlib.h:1282: error: initializer expression list treated as compound
expression
/usr/include/zlib.h:1295: error: `gzFile' was not declared in this scope
/usr/include/zlib.h:1295: error: expected primary-expression before `int'
/usr/include/zlib.h:1295: error: initializer expression list treated as compound
expression
/usr/include/zlib.h:1301: error: `gzFile' was not declared in this scope
/usr/include/zlib.h:1307: error: `gzFile' has not been declared
/usr/include/zlib.h:1319: error: `gzFile' was not declared in this scope
/usr/include/zlib.h:1319: error: expected primary-expression before `int'
/usr/include/zlib.h:1319: error: initializer expression list treated as compound
expression
/usr/include/zlib.h:1354: error: `gzFile' was not declared in this scope
/usr/include/zlib.h:1382: error: `gzFile' was not declared in this scope
/usr/include/zlib.h:1397: error: `gzFile' was not declared in this scope
/usr/include/zlib.h:1413: error: `gzFile' was not declared in this scope
/usr/include/zlib.h:1425: error: `gzFile' was not declared in this scope
/usr/include/zlib.h:1426: error: `gzFile' was not declared in this scope
/usr/include/zlib.h:1437: error: `gzFile' was not declared in this scope
/usr/include/zlib.h:1437: error: expected primary-expression before `int'
/usr/include/zlib.h:1437: error: initializer expression list treated as compound
expression
/usr/include/zlib.h:1453: error: variable or field `gzclearerr' declared void
/usr/include/zlib.h:1453: error: `gzFile' was not declared in this scope
/usr/include/zlib.h:1469: error: `uLong' does not name a type
/usr/include/zlib.h:1498: error: `uLong' does not name a type
/usr/include/zlib.h:1566: error: `gzFile' does not name a type
/usr/include/zlib.h:1567: error: `z_off64_t' does not name a type
/usr/include/zlib.h:1568: error: `z_off64_t' does not name a type
/usr/include/zlib.h:1569: error: `z_off64_t' does not name a type
/usr/include/zlib.h:1570: error: `uLong' does not name a type
/usr/include/zlib.h:1571: error: `uLong' does not name a type
/usr/include/zlib.h:1590: error: `gzFile' does not name a type
/usr/include/zlib.h:1591: error: `gzFile' was not declared in this scope
/usr/include/zlib.h:1591: error: expected primary-expression before `,' token
/usr/include/zlib.h:1591: error: expected primary-expression before `int'
/usr/include/zlib.h:1591: error: initializer expression list treated as compound
expression
/usr/include/zlib.h:1592: error: `gzFile' was not declared in this scope
/usr/include/zlib.h:1593: error: `gzFile' was not declared in this scope
/usr/include/zlib.h:1594: error: `uLong' does not name a type
/usr/include/zlib.h:1595: error: `uLong' does not name a type
/usr/include/zlib.h:1606: error: expected initializer before `*' token
In file included from /usr/include/libpng12/png.h:518,
                  from /home/sterling/OpenCV-2.1.0/src/highgui/grfmt_png.cpp:57:
/usr/include/libpng12/pngconf.h:1360: error: expected initializer before `*'
token
/usr/include/libpng12/pngconf.h:1361: error: expected initializer before `*'
token
[/code]


I have been trying to find a way to make this work, including actually going it
and trying to edit zconf.h until it compiles. Nothing has worked. I cannot find
anything on the web about it. Has anyone run into this problem before? If so,
how did you fix it? Any help would be greatly appreciated.

Reply | Messages in this Topic (1)
#87616 From: "Sylvain BERNHARDT" <sylber44@...>
Date: Thu May 3, 2012 9:16 pm
Subject: findFundamentalMatrix assertion failed 		skysonic2003
Send Email Send Email
  
Hi all,
I'm using OpenSURF to match points between two frames, and I would like to
calculate the fundamental matrix. So here is what I have so far:

int main(int argc, char **argv)
{
...
vector<Point2f> pts1(*(surf.GetPts1()));
vector<Point2f> pts2(*(surf.GetPts1()));

Mat F = findFundamentalMat(pts1, pts2);
...
}

Although the compilation works fine and (pts1,pts2) are correctly defined with
all the points (I checked in debugging), I keep getting this assertion failed
from fundam.cpp:

Assertion failed ( npoints >= 0 && points2.checkVector(2) == npoints && 
points1.type() == points2.type())

Also, I tested with the same code as in
(http://stackoverflow.com/questions/6832983/findfundamentalmatrix-doesnt-find-fu\
ndamental-matrix) and it gave me the same error.

Any idea what could be wrong ?

Regards,
S.

Reply | Messages in this Topic (2)
#87617 From: "mursalin_sir" <mursalin_sir@...>
Date: Thu May 3, 2012 10:10 pm
Subject: Trying to create a 32 image with opencv 		mursalin_sir
Send Email Send Email
  
Dear Members,
              I am trying to create a 32 bit image but the image I am getting as
output is 8bit image.  Here is the code:

IplImage *im32 = cvCreateImage(cvGetSize(objectImg), IPL_DEPTH_32F, 1);
cvSaveImage("test32.tiff", im32, NULL);

Thanks

Reply | Messages in this Topic (3)
#87618 From: "mursalin_sir" <mursalin_sir@...>
Date: Thu May 3, 2012 10:16 pm
Subject: Conversion from 16 bit image to 32 bit image 		mursalin_sir
Send Email Send Email
  
I have spent two days trying this but failed.  I am trying to convert a 16bit
image to 32 bit floating point image. Here is the code.  It gives me a 8bit
image:

IplImage *im32 = cvCreateImage(cvGetSize(objectImg), IPL_DEPTH_32F,
objectImg->nChannels);

     cvDiv(objectImg, newImg,cellImg1, 1);
     cvConvertScale(cellImg1,im32,1);
     cvSaveImage("test32.tiff", im32, NULL);

thanks

Reply | Messages in this Topic (2)
#87619 From: "alexferrier@..." <alexferrier@...>
Date: Thu May 3, 2012 10:43 pm
Subject: Re: 2.4 slower than 2.3? 		alexferrier...
Send Email Send Email
  
Sorry for not replying - I haven't looked into how to get Yahoogroups to notify
of replies.

I just fixed it manually in XCode. CMake is a complex meta-build system of the
sort you need for any serious multi-developer, multi-package project setup, but
for simple projects you can cheat and just use CMake to generate approximately
the right thing, then fix the XCodeProj up manually.

In other words, run your build script, then open the xcodeproj and delete the
GCC_OPTIMIZATION_LEVEL define in XCode project settings, shift-select all the
lib targets and set their release build optimization level to -O3 or -Os. Then I
cleaned the Release configuration, built manually, and linked directly from the
lib/Release folder.

I'd love to fix this more elegantly (and also solve my other question about why
the exported libs for Mac Static builds seem to contain incorrect framework
questions), but couldn't spend any more time investigating myself.

I also reported the issue at the base of here, and one other poster has
confirmed they see the same thing. Note that the script you're using, is I
believe derived from this one with additional steps to build a Framework.

http://computer-vision-talks.com/2011/03/opencv-build-script-with-xcode-4-suppor\
t-is-here/

Alex

--- In OpenCV@yahoogroups.com, Radu Spineanu <spineanur@...> wrote:
>
> Hey Alex,
>
> You may be right. both my build script and the build script you linked too is
compiling openCV with -O0. What did you do to fix it?
>
> -R.
>
>
> ________________________________
>  From: "alexferrier@..." <alexferrier@...>
> To: OpenCV@yahoogroups.com
> Sent: Wednesday, April 25, 2012 2:38 AM
> Subject: [OpenCV] Re: 2.4 slower than 2.3?
>
>
>  
> Not that I'm doubting you, but I just had exactly the same issue with a 2.4
upgrade, and it WAS due to release build actually having debug codeGen
(previously I'd been using a precompiled lib of 2.3.) I haven't looked in detail
at your build scripts, but with a different script, I've just posted a reply at
the footer of this page here:
>
>
http://computer-vision-talks.com/2011/03/opencv-build-script-with-xcode-4-suppor\
t-is-here/
>
> Unfortunately, if that isn't your issue, it's time to open the OpenCV
workspace, look at the CPP files where your time is mostly spent, and check
code.opencv.org for edits to those files between 2.4 and 2.3. Sometimes a good
old binary-search in release space will get you your answer in a couple of
hours...
>
> --- In OpenCV@yahoogroups.com, Radu Spineanu <spineanur@> wrote:
> >
> > It is compiled in release mode. I'm using the build script here:
> > https://github.com/aptogo/OpenCVForiPhone/blob/master/opencvbuild.sh
> >
> > -R.
> >
> >
> >
> > ________________________________
> >  From: Walter Lucetti <w.lucetti@>
> > To: OpenCV@yahoogroups.com
> > Sent: Monday, April 23, 2012 3:05 PM
> > Subject: Re: [OpenCV] 2.4 slower than 2.3?
> >
> >
> >  
> > Verify that you compiled OpenCV 2.4 in Release mode.
> >
> > On 23/04/2012 3.22, spineanur wrote:
> >  
> > >Hi!
> > >
> > >I was hunting down a bug today so I had to play with a
> >               number of openCV versions on the same code.
> > >
> > >For my use case, getting videoFrames and doing phase
> >               correlation, it seems like 2.4 is significantly slower. It
> >               takes me 1.3 seconds now to do Phase Correlation between
> >               two reasonably sized frames on an iOS device.
> > >
> > >If I can't figure out a way to improve the performance I
> >               won't be able to use it in my project anymore. Or I'll
> >               have to stick to 2.3.
> > >
> > >Is anything being done about this? Can I do something at
> >               compile time?
> > >
> > >-R.
> > >
> > >
> >
>

Reply | Messages in this Topic (7)
#87620 From: "houyang20526" <houyang20526@...>
Date: Fri May 4, 2012 12:51 am
Subject: Re: Bag of words 		houyang20526
Send Email Send Email
  
I working on the bag-of-words classification and  have few questions regarding
that.Firstly ,I used Opencv Bag of words code(bagofwords_classification.cpp) and
succesfully train visual word vocabulary , and getting good SVM classifier after
training,but the predict is always -1.Please let me know  how to resolve this
issue.Thank you!

--- In OpenCV@yahoogroups.com, "maheshg_iitm" <maheshg_iitm@...> wrote:
>
> Hi,
>
> I am Mahesh, working on object classification using OpenCV Bag of words code.
>
> I used OpenCv Bag of words code and succesfully trained for VOC Pascal
database (cluster size = 1000). Since VOC database images have lot of background
so I removed background and cropped only object class region using
annotations.xml files (given in VOC pascal database). Then I tried to train the
new object class images (cropped) for two classes people and vehicles. I am
getting zero support vectors after training. What will be the issue? Is number
clusters a problem ( I tried 1000, 500 and 50, in all cases i got zero support
vectors). Will bacground effect the results? I used SIFT for key point
extraction, SIFT for computing image descriptor and BRUTEFORCE method for
descriptor matching.
>
> I am wondering what will be the reason for getting zero support vectors. Do I
need to change any parameters in the code. Please let me know if any one has
information on how to resolve this issue.
>
> Thanks a lot
> Mahesh
>

Reply | Messages in this Topic (7)
#87621 From: iy8tyuyu ytyuiyi <sanho_cavoixanh@...>
Date: Fri May 4, 2012 3:08 am
Subject: PlateDetect_Emgu 		sanho_cavoixanh
Send Email Send Email
 
Hi.
I'm having a problem with LicensePlateRecognition Example from Emgu CV. 
I don't know what is going on, closes my application from scratch, without giving any error

Reply | Messages in this Topic (1)
#87622 From: "nagendervss" <nagendervss@...>
Date: Fri May 4, 2012 5:24 am
Subject: Re: [GPU] Differences in the descriptors of GPU and CPU implementations of SURF 		nagendervss
Send Email Send Email
  
Thank you for your reply.

I am currently trying to use SURF to describe features from a large database of
images as a part of a project. Because of the large size of the database we are
looking at, we want to process the images on heterogeneous resources (CPUs and
GPUs) to improve performance of the system.

But to achieve this usage of heterogeneous resources, we need to get same
results (both keypoints and descriptors) independent of whether we process the
image on CPU or GPU. Is there any way to achieve this using the OpenCV
implementations of SURF (CPU and GPU)?

In one of the previous replies it has been mentioned that "Achieving complete
similarity will lead to big performance penalty". I request you to give some
more details on the performance penalty in terms of speed and accuracy.

Also, are there any plans to write Java wrappers to GPU implementations like the
ones you have for CPU implementations of OpenCV?

--- In OpenCV@yahoogroups.com, "vlad.vinogradov" <vlad.vinogradov@...> wrote:
>
> Implementations differs in interpolation mode in internal resize call.
> Both implementations get patch around keypoint with size dependent on
keypoint's size and resize it to 20x20 size.
> CPU implementation uses INTER_AREA interpolation mode, which is difficult for
parallel implementation.
> GPU implementation uses bilinear interpolation.
>
> --- In OpenCV@yahoogroups.com, "nagendervss" <nagendervss@> wrote:
> >
> >
> >
> > Hello Anatoly,
> >
> > Thanks for your reply. Can you elaborate more on the "performance penalty"
part in your reply. I would like to know the major differences between the two
implementations and also the "performance penalty" that we might face if we try
to get the same descriptors as well.
> >
> > Also what are the changes that have to be made in either the CPU
implementation or the GPU implementation in order to get both the Keypoints and
their Descriptors same in both the implementations?
> >
> > Thank you in advance.
> >
> > --- In OpenCV@yahoogroups.com, "anatoly_2011" <anatolyb2011@> wrote:
> > >
> > > Descriptors are different. Achieving complete similarity will lead to big
performance penalty. Now there is a some balance. No plans to change this. So if
you have SURF descriptors database, you need to recreate it with GPU SURF.
> > >
> > > --- In OpenCV@yahoogroups.com, "nagendervss" <nagendervss@> wrote:
> > > >
> > > > Hi everyone,
> > > >
> > > > I am trying to compare the outputs of SURF when implemented on GPU and
CPU. I find that the GPU and the CPU versions give similar keypoints but there
is a variation in the descriptors of the keypoints. The following are the codes
that I used to check the implementations
> > > >
> > > > GPU
> > > >
> > > > using namespace std;
> > > > using namespace cv;
> > > > using namespace cv::gpu;
> > > >
> > > >
> > > > int main(int argc, char* argv[])
> > > > {
> > > >
> > > > GpuMat img1(imread("lena.jpg", CV_LOAD_IMAGE_GRAYSCALE));
> > > > SURF_GPU surf;
> > > >
> > > > // detecting keypoints & computing descriptors
> > > > GpuMat keypointsGPU;
> > > > GpuMat descriptorsGPU;
> > > >
> > > > surf.extended = false;
> > > > surf(img1, GpuMat(), keypointsGPU, descriptorsGPU);
> > > >
> > > > vector<KeyPoint> keypoints;
> > > > vector<float> descriptors;
> > > >
> > > > surf.downloadKeypoints(keypointsGPU, keypoints);
> > > > surf.downloadDescriptors(descriptorsGPU, descriptors);
> > > >
> > > > int k=0;
> > > > for (int i=0; i<keypointsGPU.cols; i++)
> > > > {
> > > >  cout << i << endl;
> > > >  cout << keypoints[i].pt.x << " " << keypoints[i].pt.y << " " <<
keypoints[i].class_id << " " << keypoints[i].angle << endl;
> > > >  for (int j=0; j<64; j++, k++)
> > > > 	 cout << descriptors[k] << " ";
> > > >  cout << endl;
> > > > }
> > > >
> > > > IplImage* image1 = cvLoadImage("lena.jpg",CV_LOAD_IMAGE_UNCHANGED);
> > > > int radius = 3;
> > > > for(int i=0;i<keypointsGPU.cols;i++)
> > > > 	 cvCircle(image1,cvPoint((int)cvRound(keypoints[i].pt.x),
(int)cvRound(keypoints[i].pt.y)) , radius, CV_RGB(255,0,0), 1,8,0);
> > > > cvSaveImage("lenaGPU02.jpg",image1);
> > > >
> > > > return 0;
> > > >
> > > > }
> > > >
> > > >
> > > > CPU
> > > >
> > > >
> > > > using namespace std;
> > > > using namespace cv;
> > > >
> > > > int main(int argc, char** argv)
> > > > {
> > > >
> > > > Mat image = imread("lena.jpg", CV_LOAD_IMAGE_GRAYSCALE);
> > > >
> > > > SURF surfcpu;
> > > >
> > > > surfcpu.extended = false;
> > > >
> > > > vector<KeyPoint> keypoints;
> > > > Mat descriptors;
> > > >
> > > > surfcpu(image, Mat(), keypoints, descriptors, false);
> > > >
> > > > for(int i=0; i<keypoints.size(); i++)
> > > > {
> > > >  cout<<i<<endl;
> > > >  cout<<keypoints[i].pt.x<<" "<<keypoints[i].pt.y<<"
"<<keypoints[i].class_id<<" "<<keypoints[i].angle<<endl;
> > > >  for(int j=0; j<64; j++)
> > > > 	 cout<<descriptors.at<float>(i,j)<<" ";
> > > >  cout<<endl;
> > > > }
> > > >
> > > > IplImage* image1 = cvLoadImage("lena.jpg",CV_LOAD_IMAGE_UNCHANGED);
> > > > int radius = 3;
> > > > for(int i=0;i<keypoints.size();i++)
> > > > 	 cvCircle(image1,cvPoint((int)cvRound(keypoints[i].pt.x),
(int)cvRound(keypoints[i].pt.y)) , radius, CV_RGB(0,255,0), 1,8,0);
> > > > cvSaveImage("lenaCPU02.jpg",image1);
> > > >
> > > > return 0;
> > > >
> > > > }
> > > >
> > > >
> > > > Does anyone know why the above two implementations print different
descriptors though the keypoints seem to be similar?? I am using the beta
release of OpenCV 2.4.0 with CUDA 4.1 installed on GPU (nvidia).
> > > >
> > > > Thank you.
> > > >
> > >
> >
>

Reply | Messages in this Topic (5)
#87623 From: "zyk0902" <zyk0902@...>
Date: Fri May 4, 2012 5:30 am
Subject: Convert LatentSVM Mat to XML file 		zyk0902
Send Email Send Email
  
Hi All,

I am trying to train my own latent SVM for people detection using [Felzenszwalb
2010]'s code, which is written in Matlab.

After training for dozens of hours, I finally got a .mat file for Matlab model.

Now, I want to convert this .mat file into .xml file so that I can use it in
OpenCV LatentSVM facilities. There is a VOCWriteXML function in VOCDevkit, but
it does not work in this case.

[Felzenszwalb 2010]'s Code
http://www.cs.brown.edu/~pff/latent/

VOCWriteXML
http://pascallin.ecs.soton.ac.uk/challenges/VOC/voc2010/htmldoc/devkit_doc.html#\
SECTION000106000000000000000

So, is there anyone who succeeds in converting the .mat file into a .xml file?

Thanks!

Reply | Messages in this Topic (1)
#87624 From: Andrew Keeling <andrewkeeling@...>
Date: Fri May 4, 2012 6:24 am
Subject: Re: Re: Reading / Writing MS Excel files in Opencv 		andrewkeeling@...
Send Email Send Email
 

;

Sent from my HTC

----- Reply message -----
From: "munir" <munircs@...>
To: <OpenCV@yahoogroups.com>
Subject: [OpenCV] Re: Reading / Writing MS Excel files in Opencv
Date: Thu, May 3, 2012 10:20 PM
 

Thanks a lot Kishor Durve.
yep it helps.........

--- In OpenCV@yahoogroups.com, Kishor Durve <kishor_durve@...> wrote:
>
> �I think you will have to create a .csv (comma separated) file and use a conversion utility like Dataload to convert this file to .xls -�at least this��is what I do.
>
> Kishor Durve
> 'A single dream is more powerful than a thousand realities'
>


Reply | Messages in this Topic (1)
#87625 From: Yili Zhao <panovr@...>
Date: Fri May 4, 2012 8:56 am
Subject: OpenCV add support for GTK+ 3.x 		panovr...
Send Email Send Email
  
Hi
   I tried to install OpenCV 2.4.0 on Ubuntu 12.04.

   Using cmake-gui, it said that OpenCV needs GTK+ 2.x, named as "libgtk2.0-dev".

   Because GTK+ 3.x is the mainstream toolkit version in Ubuntu 12.04
and other Linux distributions,
may OpenCV add support for GTK+ 3.x, named as "libgtk-3-dev"?

   Thanks!

--
Yili Zhao

Reply | Messages in this Topic (1)
#87626 From: "Mai" <traitim_tanvo98@...>
Date: Fri May 4, 2012 10:11 am
Subject: Re: How to get the frames from IP camera?. 		traitim_tanvo98
Send Email Send Email
  
Help me! I also have this problem. How to read data from Ip camera with OpenCV?
Please can somebody help me!!!
--- In OpenCV@yahoogroups.com, "minimanu74" <minimanu74@...> wrote:
>
> Hi Thomas,
> i have an axis too, i have tryed to use your code but it doesn't work
> for me.
> Is there any settings to do ?
> thanx
> bye
> Manuela
>
> --- In OpenCV@yahoogroups.com, "tdenewiler" <tdenewiler@> wrote:
> >
> > Depending on the IP camera you have this can be very simple. I have an
> > Axis camera that works using TCP/IP and sends out mjpeg and MPEG4
> > video streams simultaneously. I can use a web browser to view the
> > video and if I put that URL into OpenCV then I can get the frames that
> > way:
> >
> > <code>
> >  CvCapture* capture =
> > cvCaptureFromFile("http://192.168.1.25/mjpg/video.mjpg");
> > </code>
> >
> > I have not tried this with any other IP camera. There are also many
> > other options to set using the URL. Things like the resolution, frame
> > rate, etc., can be set by passing in the appropriate options on the URL.
> >
> > Hope that helps.
> >
> > Thomas
> >
> >
> >
> > --- In OpenCV@yahoogroups.com, "harryk128" <harryk128@> wrote:
> > >
> > > Hi Senthil,
> > >
> > > in my project I have an IP camera too. I didn't found a function in
> > > opencv which grabs frames from an IP camera, so I wrote my one code.
> > > How you have to do it depends on your camera and your
> > > programming language. What programming language do you use?
> > >
> > > Harry
> > >
> >
>

Reply | Messages in this Topic (6)
#87627 From: "martinsmails57" <martinsmails57@...>
Date: Fri May 4, 2012 10:05 am
Subject: Create Image Line from Column Means 		martinsmails57
Send Email Send Email
  
Hello folks,
I would like to create a function, which calculates the mean of each
column of an image, and return it again as an image.
Example: Input 200x200 Pixels, Output 1x200 Pixels. Each single column
value is the mean of the original column.

Where I get stuck is two things:
1) how do I set the data type of the output image, so that it is double
values with the appropriate channel depth at runtime?

2) setting the mean into the new image. The function shall be
independent of the used channel count (Graysacale or RGB).

I use cv::mean as function for calculating, but it returns a scalar with
channel size determined at runtime. How do I set this into a cv::Mat
data structure?

Help is appreciated. Code stub (not executable) is at the end.
Greetings from Germany. Martin

Here my code stub:

int calc_img_col_mean(cv::Mat* p_img_src, cv::Mat* p_img_line_dest) {

      //! prepare destination image @TODO Datatype
      *p_img_line_dest.create(1,*p_img_src.size.width, <double>);

      for (int i = 0; i < *p_img_line_dest.cols; i++) {
          // get source column
          cv::Mat colVal = *p_img_src.col(i);
          // calc col mean
          cv::Scalar colMean = cv::mean(colVal,cv::noArray());
          //! overwrite dest col @TODO write access
          *p_img_line_dest<DATATYPE>(1,i) = colMean;
      }
}

Reply | Messages in this Topic (1)
#87629 From: Hemant Singh <unizen01@...>
Date: Thu May 3, 2012 1:02 pm
Subject: Re: Object Detection and tracking in opencv 		maddyhemu
Send Email Send Email
 
Report your progress, what all research you have done, then we can help you better.

On Wed, May 2, 2012 at 5:54 PM, Muhammad <waqar.uet23@...> wrote:
 

hay!

I want to detect and track the motion of my foot and want to do actions accordingly. kindly help me.




--


-Hemant 


Reply | Messages in this Topic (2)
#87630 From: "anatoly_2011" <anatolyb2011@...>
Date: Fri May 4, 2012 1:13 pm
Subject: Re: Issue with GPU and GeForce GTX 680 		anatoly_2011
Send Email Send Email
  
Yup! I develop both PCL and OpenCV for GPU

--- In OpenCV@yahoogroups.com, Walter Lucetti <w.lucetti@...> wrote:
>
> Hi Anatoly,
>
> I finally reached a good configuration that let me test the card.
>
> The final result for "performance_gpu" is "average GPU speedup: x41.827"
>
> I also tested "stereo_match_gpu" with "aloe" and "tsucuba"
>
> These are the results:
>
> ALOE about 90 FPS
> --- Parameters ---
> image_size: (1282, 1110)
> image_channels: 1
> method: BM
> ndisp: 64
> win_size: 19
> prefilter_sobel: 0
>
> TSUCUBA about 740 FPS (!!!)
> --- Parameters ---
> image_size: (384, 288)
> image_channels: 1
> method: BM
> ndisp: 64
> win_size: 19
> prefilter_sobel: 0
>
>
>
>
>
> Cheers
> Walter
>
> PS Are you the same Anatoly I am talking with in PCL mailing list?
>
> 2012/5/4 anatoly_2011 <anatolyb2011@...>
>
> > **
> >
> >
> > BTW, if you add '30' to CUDA_ARCH_BIN varibable in OPenCV CMake, the code
> > is supposed to work faster.
> >
> > BTW, I haven't had any chance to test OpenCV with Kepler.
> >
> >
> > --- In OpenCV@yahoogroups.com, Walter Lucetti <w.lucetti@> wrote:
> > >
> > > It was my hurry... after waiting a few seconds the test application
> > started
> > > and finally I got it working.
> > >
> > > Thank you
> > > Walter
> > >
> > > 2012/5/3 Thomas Bastiani <thom@>
> > >
> > > > **
> >
> > > >
> > > >
> > > > Hi Walter,
> > > >
> > > > Can you give us the exact revision of your CUDA Driver, toolkit and
> > your
> > > > CUDA SDK?
> > > > Also, can you try running the deviceQuery CUDA sample to see if it
> > > > completes?
> > > >
> > > > Eventually, if deviceQuery freezes up, build it in debug mode and run
> > it
> > > > under gdb so that we know where it's blocked.
> > > >
> > > > Cheers,
> > > > --
> > > > Thomas
> > > >
> > > >
> > > > On 3 May 2012 09:46, Walter Lucetti <w.lucetti@> wrote:
> > > >
> > > >>
> > > >>
> > > >> Hi Anatoly,
> > > >>
> > > >> I updated to latest trunk (rev. 8295) and now Kepler device info is
> > > >> printed correctly.
> > > >>
> > > >> I tried again "performance_gpu" example and it freezes at this point:
> > > >>
> > > >> ======
> > > >> Device 0: "GeForce GTX 680" 2048Mb, sm_30, 1536 cores, Driver/Runtime
> > > >> ver.4.20/4.20
> > > >>
> > > >> CPU, ms GPU, ms SPEEDUP DESCRIPTION
> > > >>
> > > >> matchTemplate:
> > > >> ======
> > > >>
> > > >> Each gpu test application seems to freeze at the beginning of
> > computation.
> > > >>
> > > >> Walter
> > > >>
> > > >>
> > > >> 2012/5/3 Walter Lucetti <w.lucetti@>
> >
> > > >>
> > > >>> Thank you Anatoly. Tomorrow I will take the latest trunk instead of
> > the
> > > >>> stable.
> > > >>>
> > > >>> This morning (Italy time) it didn't compile and I went straight to
> > > >>> official version.
> > > >>>
> > > >>>
> > > >>> 2012/5/2 anatoly_2011 <anatolyb2011@>
> > > >>>
> > > >>>> **
> >
> > > >>>>
> > > >>>>
> > > >>>> This is fixed in trunk version in r8293
> > > >>>>
> > > >>>>
> > > >>>> --- In OpenCV@yahoogroups.com, Walter Lucetti <w.lucetti@> wrote:
> > > >>>> >
> > > >>>> > Hi all,
> > > >>>> >
> > > >>>> > I installed new OpenCV 2.4 and compiled with latest CUDA driver,
> > > >>>> toolkit
> > > >>>> > and SDK under Linux 11.10.
> > > >>>> >
> > > >>>> > I compiled everything without any problems, but I cannot run
> > examples
> > > >>>> and
> > > >>>> > test (they all freeze).
> > > >>>> >
> > > >>>> > The only example that gives a feedback is "performance_gpu" and I
> > get
> > > >>>> this
> > > >>>> > error:
> > > >>>> >
> > > >>>> > "Device 0: "GeForce GTX 680" 2048MbMapSMtoCores undefined
> > SMversion
> > > >>>> 3.0!"
> > > >>>> >
> > > >>>> > Any idea?
> > > >>>> >
> > > >>>> > Thank you
> > > >>>> > Walter
> > > >>>> >
> > > >>>>
> > > >>>>
> > > >>>
> > > >>
> > > >>
> > > >>
> > > >
> > > >
> > >
> >
> >
> >
>

Reply | Messages in this Topic (9)
#87631 From: Vaclav <vaclav_sal@...>
Date: Fri May 4, 2012 2:06 pm
Subject: Re: For Help in too few arguments to function ‘cvHaarDetectObjects’ 		vaclav_sal
Send Email Send Email
 
A compiler "error".
It suppose to do that when the function declaration (parameters count ) and usage do not match.
 
Temporary delete the offending function execution and run your code in debug mode to find out which parameter is missing or dummy up all the paramaters, all 8 of them,  and step to the offending function.
 
You could also change the function declaration and add default paramaters values so the compiler is happy.
 
Ether way, posting a code snippet here would help to help you in the future.
Vaclav

Reply | Messages in this Topic (3)
Messages 87601 - 87631 of 90774   Oldest  |  < Older  |  Newer >  |  Newest
Add to My Yahoo!      XML What's This?
Copyright © 2010 Yahoo! Inc. All rights reserved.
Privacy Policy - Terms of Service - Guidelines NEW - Help