<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"
	xmlns:content="http://purl.org/rss/1.0/modules/content/"
	xmlns:dc="http://purl.org/dc/elements/1.1/"
	xmlns:atom="http://www.w3.org/2005/Atom"
	xmlns:sy="http://purl.org/rss/1.0/modules/syndication/"
		>
<channel>
	<title>Comments on: FAQ: OpenCV Haartraining</title>
	<atom:link href="http://www.computer-vision-software.com/blog/2009/11/faq-opencv-haartraining/feed/" rel="self" type="application/rss+xml" />
	<link>http://www.computer-vision-software.com/blog/2009/11/faq-opencv-haartraining/</link>
	<description>Rhonda Ltd., computer vision software blog</description>
	<lastBuildDate>Mon, 06 Sep 2010 10:44:54 +0000</lastBuildDate>
	<sy:updatePeriod>hourly</sy:updatePeriod>
	<sy:updateFrequency>1</sy:updateFrequency>
	<generator>http://wordpress.org/?v=3.0.1</generator>
	<item>
		<title>By: Andrey Soldatov</title>
		<link>http://www.computer-vision-software.com/blog/2009/11/faq-opencv-haartraining/comment-page-1/#comment-809</link>
		<dc:creator>Andrey Soldatov</dc:creator>
		<pubDate>Mon, 05 Jul 2010 08:13:45 +0000</pubDate>
		<guid isPermaLink="false">http://www.computer-vision-software.com/blog/?p=106#comment-809</guid>
		<description>Unfortunately, you are right, cv::CascadeClassifier with cascade trained by traincascade doesn&#039;t support these flags. As sample cascades provided by opencv are in &quot;old&quot; format, I believe it&#039;s better to use haartraining with cvHaarDetectObjects or cv::HaarClassifierCascade::detectMultiScale (its OOP form) if you needn&#039;t new features of &quot;new&quot; cascades.</description>
		<content:encoded><![CDATA[<p>Unfortunately, you are right, cv::CascadeClassifier with cascade trained by traincascade doesn&#8217;t support these flags. As sample cascades provided by opencv are in &#8220;old&#8221; format, I believe it&#8217;s better to use haartraining with cvHaarDetectObjects or cv::HaarClassifierCascade::detectMultiScale (its OOP form) if you needn&#8217;t new features of &#8220;new&#8221; cascades.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: jraul</title>
		<link>http://www.computer-vision-software.com/blog/2009/11/faq-opencv-haartraining/comment-page-1/#comment-799</link>
		<dc:creator>jraul</dc:creator>
		<pubDate>Tue, 22 Jun 2010 19:24:25 +0000</pubDate>
		<guid isPermaLink="false">http://www.computer-vision-software.com/blog/?p=106#comment-799</guid>
		<description>Hi all!!!

I implemented a face recognition system based on OpenCV HaarCascadeClassifier and eigenfaces. Now I want to review the code and use new functionality of OpenCV. 

I start to read and I found the next text:

flags – This parameter is not used for new cascade and have the same meaning for old cascade as in function cvHaarDetectObjects.

It&#039;s mean that cascades trained with the new traincascade app doesn`t supprot this flags? Some of them are very usefuls sometimes.

Thanks for your answers!</description>
		<content:encoded><![CDATA[<p>Hi all!!!</p>
<p>I implemented a face recognition system based on OpenCV HaarCascadeClassifier and eigenfaces. Now I want to review the code and use new functionality of OpenCV. </p>
<p>I start to read and I found the next text:</p>
<p>flags – This parameter is not used for new cascade and have the same meaning for old cascade as in function cvHaarDetectObjects.</p>
<p>It&#8217;s mean that cascades trained with the new traincascade app doesn`t supprot this flags? Some of them are very usefuls sometimes.</p>
<p>Thanks for your answers!</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: sbrovche</title>
		<link>http://www.computer-vision-software.com/blog/2009/11/faq-opencv-haartraining/comment-page-1/#comment-777</link>
		<dc:creator>sbrovche</dc:creator>
		<pubDate>Sat, 12 Jun 2010 05:18:56 +0000</pubDate>
		<guid isPermaLink="false">http://www.computer-vision-software.com/blog/?p=106#comment-777</guid>
		<description>Bob,

I suspect you encounted the issue with non-initialization of static haar_type variable inside harr.cpp. During initialization of haar_type variable, some important information is registered, that is used later to help parse XML file during cvLoad call. Depending on OpenCV version, operation system and static vs dynamic library linkage, it might happen that this static variable is not initialized before cvLoad is invoked, and then you may get this error.
To resolve the issue, you need to invoke some function in object detection library *before* invoking cvLoad for haar cascade file. This forces initialization of all static variables including haar_type.
cvReleaseHaarClassifierCascade with pointer to empty cascade worked fine for me.
...
CvHaarClassifierCascade * cascade = 0;
cvReleaseHaarClassifierCascade(&amp;cascade);
cascade = (CvHaarClassifierCascade) cvLoad(path);
...

Best regards,
Sergey</description>
		<content:encoded><![CDATA[<p>Bob,</p>
<p>I suspect you encounted the issue with non-initialization of static haar_type variable inside harr.cpp. During initialization of haar_type variable, some important information is registered, that is used later to help parse XML file during cvLoad call. Depending on OpenCV version, operation system and static vs dynamic library linkage, it might happen that this static variable is not initialized before cvLoad is invoked, and then you may get this error.<br />
To resolve the issue, you need to invoke some function in object detection library *before* invoking cvLoad for haar cascade file. This forces initialization of all static variables including haar_type.<br />
cvReleaseHaarClassifierCascade with pointer to empty cascade worked fine for me.<br />
&#8230;<br />
CvHaarClassifierCascade * cascade = 0;<br />
cvReleaseHaarClassifierCascade(&amp;cascade);<br />
cascade = (CvHaarClassifierCascade) cvLoad(path);<br />
&#8230;</p>
<p>Best regards,<br />
Sergey</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Andrey Soldatov</title>
		<link>http://www.computer-vision-software.com/blog/2009/11/faq-opencv-haartraining/comment-page-1/#comment-772</link>
		<dc:creator>Andrey Soldatov</dc:creator>
		<pubDate>Fri, 11 Jun 2010 04:38:51 +0000</pubDate>
		<guid isPermaLink="false">http://www.computer-vision-software.com/blog/?p=106#comment-772</guid>
		<description>Could you clarify, do you use sample program from OpenCV (haartraining.exe for example) or write your own program. Please, provide command line in 1st case and part of your code with call of cvRead or cvLoad or .load and declaration of its arguments in 2nd case.</description>
		<content:encoded><![CDATA[<p>Could you clarify, do you use sample program from OpenCV (haartraining.exe for example) or write your own program. Please, provide command line in 1st case and part of your code with call of cvRead or cvLoad or .load and declaration of its arguments in 2nd case.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: bob</title>
		<link>http://www.computer-vision-software.com/blog/2009/11/faq-opencv-haartraining/comment-page-1/#comment-752</link>
		<dc:creator>bob</dc:creator>
		<pubDate>Thu, 27 May 2010 18:50:58 +0000</pubDate>
		<guid isPermaLink="false">http://www.computer-vision-software.com/blog/?p=106#comment-752</guid>
		<description>I  got an OpenCV GUI Error message … “Unspecified error (The node does not represent a user object (unknown type?) in function cvRead, .\cxpersistence.cpp(5061)”
how can i get rid of this problem plz ??+ Ididn&#039;t call cvread before</description>
		<content:encoded><![CDATA[<p>I  got an OpenCV GUI Error message … “Unspecified error (The node does not represent a user object (unknown type?) in function cvRead, .\cxpersistence.cpp(5061)”<br />
how can i get rid of this problem plz ??+ Ididn&#8217;t call cvread before</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Henner</title>
		<link>http://www.computer-vision-software.com/blog/2009/11/faq-opencv-haartraining/comment-page-1/#comment-751</link>
		<dc:creator>Henner</dc:creator>
		<pubDate>Thu, 27 May 2010 14:28:23 +0000</pubDate>
		<guid isPermaLink="false">http://www.computer-vision-software.com/blog/?p=106#comment-751</guid>
		<description>1. 
It seems to be a bug ... see these links
http://staff.science.uva.nl/~rvalenti/index.php?content=projects
http://opencv-users.1802565.n2.nabble.com/face-detection-cvLoad-td4467872.html

I tried to create a image and release it afterwords, but it did work =(</description>
		<content:encoded><![CDATA[<p>1.<br />
It seems to be a bug &#8230; see these links<br />
<a href="http://staff.science.uva.nl/~rvalenti/index.php?content=projects" rel="nofollow">http://staff.science.uva.nl/~rvalenti/index.php?content=projects</a><br />
<a href="http://opencv-users.1802565.n2.nabble.com/face-detection-cvLoad-td4467872.html" rel="nofollow">http://opencv-users.1802565.n2.nabble.com/face-detection-cvLoad-td4467872.html</a></p>
<p>I tried to create a image and release it afterwords, but it did work =(</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Henner</title>
		<link>http://www.computer-vision-software.com/blog/2009/11/faq-opencv-haartraining/comment-page-1/#comment-750</link>
		<dc:creator>Henner</dc:creator>
		<pubDate>Thu, 27 May 2010 14:24:52 +0000</pubDate>
		<guid isPermaLink="false">http://www.computer-vision-software.com/blog/?p=106#comment-750</guid>
		<description>2. 
facedetect.exe --cascade=&quot;haarcascade.xml&quot; 

I got it and it works ^^</description>
		<content:encoded><![CDATA[<p>2.<br />
facedetect.exe &#8211;cascade=&#8221;haarcascade.xml&#8221; </p>
<p>I got it and it works ^^</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Henner</title>
		<link>http://www.computer-vision-software.com/blog/2009/11/faq-opencv-haartraining/comment-page-1/#comment-749</link>
		<dc:creator>Henner</dc:creator>
		<pubDate>Wed, 26 May 2010 08:36:50 +0000</pubDate>
		<guid isPermaLink="false">http://www.computer-vision-software.com/blog/?p=106#comment-749</guid>
		<description>Normaly the program ends and creates the output haarcascade.xml file, if it oversteps the false alarm rate... It seems that the program can&#039;t create this haarcascade.xml file. 

Your directory structure should loks like this:
.../haarcascade/
       0/AdaBoostCARTHaarClassifier.txt
       1/AdaBoostCARTHaarClassifier.txt
       2/AdaBoostCARTHaarClassifier.txt
       3/AdaBoostCARTHaarClassifier.txt
       4/AdaBoostCARTHaarClassifier.txt
       5/AdaBoostCARTHaarClassifier.txt
       6/AdaBoostCARTHaarClassifier.txt
       7/AdaBoostCARTHaarClassifier.txt
       8/AdaBoostCARTHaarClassifier.txt
       9/AdaBoostCARTHaarClassifier.txt
     10/AdaBoostCARTHaarClassifier.txt
     11/AdaBoostCARTHaarClassifier.txt
     12/AdaBoostCARTHaarClassifier.txt
     13/AdaBoostCARTHaarClassifier.txt
     14/AdaBoostCARTHaarClassifier.txt
     15/AdaBoostCARTHaarClassifier.txt
.../haarcascade.xml

http://opencv.sourcearchive.com/documentation/1.0.0-6.1/cxpersistence_8cpp-source.html</description>
		<content:encoded><![CDATA[<p>Normaly the program ends and creates the output haarcascade.xml file, if it oversteps the false alarm rate&#8230; It seems that the program can&#8217;t create this haarcascade.xml file. </p>
<p>Your directory structure should loks like this:<br />
&#8230;/haarcascade/<br />
       0/AdaBoostCARTHaarClassifier.txt<br />
       1/AdaBoostCARTHaarClassifier.txt<br />
       2/AdaBoostCARTHaarClassifier.txt<br />
       3/AdaBoostCARTHaarClassifier.txt<br />
       4/AdaBoostCARTHaarClassifier.txt<br />
       5/AdaBoostCARTHaarClassifier.txt<br />
       6/AdaBoostCARTHaarClassifier.txt<br />
       7/AdaBoostCARTHaarClassifier.txt<br />
       8/AdaBoostCARTHaarClassifier.txt<br />
       9/AdaBoostCARTHaarClassifier.txt<br />
     10/AdaBoostCARTHaarClassifier.txt<br />
     11/AdaBoostCARTHaarClassifier.txt<br />
     12/AdaBoostCARTHaarClassifier.txt<br />
     13/AdaBoostCARTHaarClassifier.txt<br />
     14/AdaBoostCARTHaarClassifier.txt<br />
     15/AdaBoostCARTHaarClassifier.txt<br />
&#8230;/haarcascade.xml</p>
<p><a href="http://opencv.sourcearchive.com/documentation/1.0.0-6.1/cxpersistence_8cpp-source.html" rel="nofollow">http://opencv.sourcearchive.com/documentation/1.0.0-6.1/cxpersistence_8cpp-source.html</a></p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Henner</title>
		<link>http://www.computer-vision-software.com/blog/2009/11/faq-opencv-haartraining/comment-page-1/#comment-748</link>
		<dc:creator>Henner</dc:creator>
		<pubDate>Wed, 26 May 2010 08:01:00 +0000</pubDate>
		<guid isPermaLink="false">http://www.computer-vision-software.com/blog/?p=106#comment-748</guid>
		<description>1.
Hey Andrey, I tried to solve the problem with creating an image and release it afterwards, but it doesn&#039;t work:
    IplImage* temp  = cvCreateImage(cvSize(8,8), IPL_DEPTH_8U, 1);
    cvReleaseImage(&amp;temp);

It seems to be a bug (see these links):
http://www.openframeworks.cc/forum/viewtopic.php?f=10&amp;t=1853
http://opencv-users.1802565.n2.nabble.com/face-detection-cvLoad-td4467872.html

The error appears by loading the cascade:
    CvHaarClassifierCascade *cascade;
    char      *cascadeName = &quot;haarcascade.xml&quot;;
    
    // load the classifier
    cascade = ( CvHaarClassifierCascade* )cvLoad( cascadeName, 0, 0, 0 );
 
    // initialize memory buffer for the face detector 
    storage = cvCreateMemStorage( 0 ); 
cvLoads loads object from file. It provides a simple interface to cvRead. After object is loaded, the file storage is closed and all the temporary buffers are deleted. Thus, to load a dynamic structure, such as sequence, contour or graph, one should pass a valid destination memory storage to the function.

p.s. What means FileNode? :[



2.
I don&#039;t use any command, because the console close after compiling the program.</description>
		<content:encoded><![CDATA[<p>1.<br />
Hey Andrey, I tried to solve the problem with creating an image and release it afterwards, but it doesn&#8217;t work:<br />
    IplImage* temp  = cvCreateImage(cvSize(8,8), IPL_DEPTH_8U, 1);<br />
    cvReleaseImage(&amp;temp);</p>
<p>It seems to be a bug (see these links):<br />
<a href="http://www.openframeworks.cc/forum/viewtopic.php?f=10&amp;t=1853" rel="nofollow">http://www.openframeworks.cc/forum/viewtopic.php?f=10&amp;t=1853</a><br />
<a href="http://opencv-users.1802565.n2.nabble.com/face-detection-cvLoad-td4467872.html" rel="nofollow">http://opencv-users.1802565.n2.nabble.com/face-detection-cvLoad-td4467872.html</a></p>
<p>The error appears by loading the cascade:<br />
    CvHaarClassifierCascade *cascade;<br />
    char      *cascadeName = &#8220;haarcascade.xml&#8221;;</p>
<p>    // load the classifier<br />
    cascade = ( CvHaarClassifierCascade* )cvLoad( cascadeName, 0, 0, 0 );</p>
<p>    // initialize memory buffer for the face detector<br />
    storage = cvCreateMemStorage( 0 );<br />
cvLoads loads object from file. It provides a simple interface to cvRead. After object is loaded, the file storage is closed and all the temporary buffers are deleted. Thus, to load a dynamic structure, such as sequence, contour or graph, one should pass a valid destination memory storage to the function.</p>
<p>p.s. What means FileNode? :[</p>
<p>2.<br />
I don&#8217;t use any command, because the console close after compiling the program.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Yasir</title>
		<link>http://www.computer-vision-software.com/blog/2009/11/faq-opencv-haartraining/comment-page-1/#comment-745</link>
		<dc:creator>Yasir</dc:creator>
		<pubDate>Wed, 26 May 2010 06:41:05 +0000</pubDate>
		<guid isPermaLink="false">http://www.computer-vision-software.com/blog/?p=106#comment-745</guid>
		<description>Hello!

I am working on haar training of different hand gestures............ how multiple cascades are loaded into the system simultaneously, and each cascade is responsible for a specific hand posture. Is there anyway to do it........... plz help me out in this 

                        regards</description>
		<content:encoded><![CDATA[<p>Hello!</p>
<p>I am working on haar training of different hand gestures&#8230;&#8230;&#8230;&#8230; how multiple cascades are loaded into the system simultaneously, and each cascade is responsible for a specific hand posture. Is there anyway to do it&#8230;&#8230;&#8230;.. plz help me out in this </p>
<p>                        regards</p>
]]></content:encoded>
	</item>
</channel>
</rss>
