<?xml version="1.0" encoding="ISO-8859-1" ?><rss version="2.0">
<channel>
    <title>RoboRealm Forum</title>
    <link>http://www.RoboRealm.com/</link>
    <description>The newest forum threads.</description>
    <lastBuildDate>Tue, 30 Nov 1999 00:11:00 EST</lastBuildDate>
    <language>en-us</language>
    <copyright>Copyright: (C) RoboRealm, http://www.RoboRealm.com/</copyright>
    <docs>http://www.RoboRealm.com/</docs>

    <item>
        <title>Building a system to match faces to a database</title>
        <description>
        Hi all,&lt;br&gt;&lt;br&gt;I want to build a face matching system using roborealm. I need the program to do the following.&lt;br&gt;1. Monitor 4 IP cameras. These cameras are from a company called Dahua.&lt;br&gt;2. Whenever the program finds a face, it should check its library of faces. if a match is found, the name matching the face, and the date and time it was detected should be recorded in a log. If there is no match, then that fact should be recorded too.&lt;br&gt;3. If a match is found, then the matching name should be spoken.&lt;br&gt;&lt;br&gt;I have tested the face recognition module which does detect the presence of faces using my USB webcam. I am unable to find a way to interface the dahua cameras with roborealm. These cameras can talk to VLC media player via streaming though. In addition, how do I handle the name matching? I suspect I will have to use the object recognition module.&lt;br&gt;&lt;br&gt;My database consists of images of faces. Each file contains a single face. The file name is the matches with that of the person the face belongs to.&lt;br&gt;&lt;br&gt;The idea behind the above program is to create a security system for myself. I am blind and existing artificial vision techmologies do not allow me to recognise faces. Yes, I doubt the system will be 100% reliable but this is an experiment.        </description>
        <link>http://www.roborealm.com/forum/index.php?thread_id=4880</link>
        <guid>http://www.roborealm.com/forum/index.php?thread_id=4880</guid>
        <pubDate>Tue, 30 Nov 1999 00:11:00 EST</pubDate>
    </item>
</channel>
</rss>
