<?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: Camshift: smile you are tracking!</title>
	<atom:link href="http://www.mukimuki.fr/flashblog/2009/05/28/camshift-smile-you-are-tracking/feed/" rel="self" type="application/rss+xml" />
	<link>http://www.mukimuki.fr/flashblog/2009/05/28/camshift-smile-you-are-tracking/</link>
	<description>The Actionscript Wrestler !</description>
	<lastBuildDate>Tue, 12 Jul 2011 07:48:40 +0200</lastBuildDate>
	<generator>http://wordpress.org/?v=2.8</generator>
	<sy:updatePeriod>hourly</sy:updatePeriod>
	<sy:updateFrequency>1</sy:updateFrequency>
		<item>
		<title>By: Og2t</title>
		<link>http://www.mukimuki.fr/flashblog/2009/05/28/camshift-smile-you-are-tracking/comment-page-1/#comment-1421</link>
		<dc:creator>Og2t</dc:creator>
		<pubDate>Fri, 23 Apr 2010 11:42:01 +0000</pubDate>
		<guid isPermaLink="false">http://www.mukimuki.fr/flashblog/?p=235#comment-1421</guid>
		<description>Forgot to say I did the haar+camshift experiment (http://play.blog2t.net/files/black-or-white/faceit.html) and it works but it&#039;s veery dependent on the background. I am thinking of trying another approach, by narrowing the camshift area to the head area only.

@Muki Muki
Have you been working on anything interesting lately?</description>
		<content:encoded><![CDATA[<p>Forgot to say I did the haar+camshift experiment (<a href="http://play.blog2t.net/files/black-or-white/faceit.html" rel="nofollow">http://play.blog2t.net/files/black-or-white/faceit.html</a>) and it works but it&#8217;s veery dependent on the background. I am thinking of trying another approach, by narrowing the camshift area to the head area only.</p>
<p>@Muki Muki<br />
Have you been working on anything interesting lately?</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Og2t</title>
		<link>http://www.mukimuki.fr/flashblog/2009/05/28/camshift-smile-you-are-tracking/comment-page-1/#comment-204</link>
		<dc:creator>Og2t</dc:creator>
		<pubDate>Mon, 03 Aug 2009 12:28:18 +0000</pubDate>
		<guid isPermaLink="false">http://www.mukimuki.fr/flashblog/?p=235#comment-204</guid>
		<description>Nice one, I am going to implement this to try out with MJ&#039;s Black and white head morphing, real acid test for CamShift ;)
You can see the current result at http://play.blog2t.net/realtime-as3-face-and-eye-detection-with-michael-jackson/</description>
		<content:encoded><![CDATA[<p>Nice one, I am going to implement this to try out with MJ&#8217;s Black and white head morphing, real acid test for CamShift <img src='http://www.mukimuki.fr/flashblog/wp-includes/images/smilies/icon_wink.gif' alt=';)' class='wp-smiley' /><br />
You can see the current result at <a href="http://play.blog2t.net/realtime-as3-face-and-eye-detection-with-michael-jackson/" rel="nofollow">http://play.blog2t.net/realtime-as3-face-and-eye-detection-with-michael-jackson/</a></p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Benjamin</title>
		<link>http://www.mukimuki.fr/flashblog/2009/05/28/camshift-smile-you-are-tracking/comment-page-1/#comment-104</link>
		<dc:creator>Benjamin</dc:creator>
		<pubDate>Tue, 14 Jul 2009 09:45:02 +0000</pubDate>
		<guid isPermaLink="false">http://www.mukimuki.fr/flashblog/?p=235#comment-104</guid>
		<description>@Merlin: You are true somehow camshift is optimized for face detection and tracking because skin color usually are really different of background colors and so it can be easily differentiated with a color detection algorithm. Furthermore the tracking window size in the original algorithm (not in my implementation)) is adapted to track an elliptical shape. I tried with some objects but it&#039;s true it does not work as well with faces. I think it is also due to the fact objects are really small into the image and I had to do some optimizations which cause an important loss of accurancy. I have some ideas to try to improve accurency but I do not have time for the moment :)</description>
		<content:encoded><![CDATA[<p>@Merlin: You are true somehow camshift is optimized for face detection and tracking because skin color usually are really different of background colors and so it can be easily differentiated with a color detection algorithm. Furthermore the tracking window size in the original algorithm (not in my implementation)) is adapted to track an elliptical shape. I tried with some objects but it&#8217;s true it does not work as well with faces. I think it is also due to the fact objects are really small into the image and I had to do some optimizations which cause an important loss of accurancy. I have some ideas to try to improve accurency but I do not have time for the moment <img src='http://www.mukimuki.fr/flashblog/wp-includes/images/smilies/icon_smile.gif' alt=':)' class='wp-smiley' /> </p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Merlin</title>
		<link>http://www.mukimuki.fr/flashblog/2009/05/28/camshift-smile-you-are-tracking/comment-page-1/#comment-97</link>
		<dc:creator>Merlin</dc:creator>
		<pubDate>Mon, 13 Jul 2009 13:30:28 +0000</pubDate>
		<guid isPermaLink="false">http://www.mukimuki.fr/flashblog/?p=235#comment-97</guid>
		<description>Excelent stuff.

Did you optimize the demo for faces somehow? I tried tracking other kinds of objects, but it failed.</description>
		<content:encoded><![CDATA[<p>Excelent stuff.</p>
<p>Did you optimize the demo for faces somehow? I tried tracking other kinds of objects, but it failed.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Trozos de Código &#124; Face detection</title>
		<link>http://www.mukimuki.fr/flashblog/2009/05/28/camshift-smile-you-are-tracking/comment-page-1/#comment-62</link>
		<dc:creator>Trozos de Código &#124; Face detection</dc:creator>
		<pubDate>Sat, 27 Jun 2009 03:22:08 +0000</pubDate>
		<guid isPermaLink="false">http://www.mukimuki.fr/flashblog/?p=235#comment-62</guid>
		<description>[...] Proyecto Spark (una implementación de Squidder y otra de Quasimodo) implementa Haar, mientras que FaceIt, también en Spark, implementa Camshift (PDF).   [...]</description>
		<content:encoded><![CDATA[<p>[...] Proyecto Spark (una implementación de Squidder y otra de Quasimodo) implementa Haar, mientras que FaceIt, también en Spark, implementa Camshift (PDF).   [...]</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Face tracking! / CODEISPOETRY</title>
		<link>http://www.mukimuki.fr/flashblog/2009/05/28/camshift-smile-you-are-tracking/comment-page-1/#comment-61</link>
		<dc:creator>Face tracking! / CODEISPOETRY</dc:creator>
		<pubDate>Fri, 26 Jun 2009 14:27:15 +0000</pubDate>
		<guid isPermaLink="false">http://www.mukimuki.fr/flashblog/?p=235#comment-61</guid>
		<description>[...] I found http://www.mukimuki.fr/flashblog/2009/05/28/camshift-smile-you-are-tracking/ [...]</description>
		<content:encoded><![CDATA[<p>[...] I found <a href="http://www.mukimuki.fr/flashblog/2009/05/28/camshift-smile-you-are-tracking/" rel="nofollow">http://www.mukimuki.fr/flashblog/2009/05/28/camshift-smile-you-are-tracking/</a> [...]</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Muki Muki » Camshift: Source and explanation</title>
		<link>http://www.mukimuki.fr/flashblog/2009/05/28/camshift-smile-you-are-tracking/comment-page-1/#comment-44</link>
		<dc:creator>Muki Muki » Camshift: Source and explanation</dc:creator>
		<pubDate>Thu, 18 Jun 2009 08:31:52 +0000</pubDate>
		<guid isPermaLink="false">http://www.mukimuki.fr/flashblog/?p=235#comment-44</guid>
		<description>[...] &#171;&#160;Camshift: smile you are tracking! [...]</description>
		<content:encoded><![CDATA[<p>[...] &laquo;&nbsp;Camshift: smile you are tracking! [...]</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Benjamin</title>
		<link>http://www.mukimuki.fr/flashblog/2009/05/28/camshift-smile-you-are-tracking/comment-page-1/#comment-25</link>
		<dc:creator>Benjamin</dc:creator>
		<pubDate>Fri, 05 Jun 2009 08:44:18 +0000</pubDate>
		<guid isPermaLink="false">http://www.mukimuki.fr/flashblog/?p=235#comment-25</guid>
		<description>@julapy: I am not an expert in computer vision but for me haar method is more used for object detection and camshift is rather for tracking. Mario aka Quasimondo had the good idea to use haar detection for face tracking but I think it&#039;s not the most efficient method for that, especially for Flash. Camshift works differently. It does not search a face but a probablity distribution into the image. That&#039;s why you do not really lose the tracking object when it leaves a bit the image and comes back. But you can use at the beginning  haar method to find the face and then following it by camshift. Well I will try to post this week-end more explanations cause some people asked me and I will try to improve it a bit because it is sometimes bugged ! :-)

@simurai: thanks! Haha unfortunately (or fortunately :-)) it is not me !</description>
		<content:encoded><![CDATA[<p>@julapy: I am not an expert in computer vision but for me haar method is more used for object detection and camshift is rather for tracking. Mario aka Quasimondo had the good idea to use haar detection for face tracking but I think it&#8217;s not the most efficient method for that, especially for Flash. Camshift works differently. It does not search a face but a probablity distribution into the image. That&#8217;s why you do not really lose the tracking object when it leaves a bit the image and comes back. But you can use at the beginning  haar method to find the face and then following it by camshift. Well I will try to post this week-end more explanations cause some people asked me and I will try to improve it a bit because it is sometimes bugged ! <img src='http://www.mukimuki.fr/flashblog/wp-includes/images/smilies/icon_smile.gif' alt=':-)' class='wp-smiley' /> </p>
<p>@simurai: thanks! Haha unfortunately (or fortunately <img src='http://www.mukimuki.fr/flashblog/wp-includes/images/smilies/icon_smile.gif' alt=':-)' class='wp-smiley' /> ) it is not me !</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: simurai</title>
		<link>http://www.mukimuki.fr/flashblog/2009/05/28/camshift-smile-you-are-tracking/comment-page-1/#comment-21</link>
		<dc:creator>simurai</dc:creator>
		<pubDate>Fri, 05 Jun 2009 07:16:25 +0000</pubDate>
		<guid isPermaLink="false">http://www.mukimuki.fr/flashblog/?p=235#comment-21</guid>
		<description>Nice! First I thought that guy would be you.. haha.. but then I remembered seeing him on youtube.</description>
		<content:encoded><![CDATA[<p>Nice! First I thought that guy would be you.. haha.. but then I remembered seeing him on youtube.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: julapy</title>
		<link>http://www.mukimuki.fr/flashblog/2009/05/28/camshift-smile-you-are-tracking/comment-page-1/#comment-20</link>
		<dc:creator>julapy</dc:creator>
		<pubDate>Fri, 05 Jun 2009 02:51:21 +0000</pubDate>
		<guid isPermaLink="false">http://www.mukimuki.fr/flashblog/?p=235#comment-20</guid>
		<description>thats interesting.
great adventage of this over haar tracking is that it copes well with tilting of the head.
haar tracking seems to lose the face when that happens.</description>
		<content:encoded><![CDATA[<p>thats interesting.<br />
great adventage of this over haar tracking is that it copes well with tilting of the head.<br />
haar tracking seems to lose the face when that happens.</p>
]]></content:encoded>
	</item>
</channel>
</rss>

