<div dir="ltr"><div><div>Hi everyone,<br><br></div>I am currently trying to use openslide to automatically open up nanozoomer ndpi files and extract regions automatically. NDP annotation files (ndpa) are used to mark ROIs. I know that annotations are not part of openslide as of yet, but I wonder if anyone has an idea about my simple query.<br></div><br>The NDP annotations seems to use their own coordinate system which I can not convert to the openslide pixel based system to extract the pictures.<br> (some of the point coordinates are actually negative!! weird! see example below) <br><br><div>    &lt;ndpviewstate id=&quot;1&quot;&gt;<br>        &lt;title&gt;area 3&lt;/title&gt;<br>        &lt;details&gt;&lt;/details&gt;<br>        &lt;x&gt;-2045952&lt;/x&gt;<br>        &lt;y&gt;2720248&lt;/y&gt;<br>        &lt;z&gt;0&lt;/z&gt;<br>        &lt;lens&gt;10.000000&lt;/lens&gt;<br>        &lt;annotation type=&quot;freehand&quot; displayname=&quot;Rectangular Area&quot; color=&quot;#000000&quot;&gt;<br>            &lt;measuretype&gt;2&lt;/measuretype&gt;<br>            &lt;closed&gt;1&lt;/closed&gt;<br>            &lt;specialtype&gt;rectangle&lt;/specialtype&gt;<br>            &lt;pointlist&gt;<br>                &lt;point&gt;<br>                    &lt;x&gt;-1999636&lt;/x&gt;<br>                    &lt;y&gt;2598706&lt;/y&gt;<br>                &lt;/point&gt;<br>                &lt;point&gt;<br>                    &lt;x&gt;-1780252&lt;/x&gt;<br>                    &lt;y&gt;2678827&lt;/y&gt;<br>                &lt;/point&gt;<br>                &lt;point&gt;<br>                    &lt;x&gt;-1831731&lt;/x&gt;<br>                    &lt;y&gt;2819784&lt;/y&gt;<br>                &lt;/point&gt;<br>                &lt;point&gt;<br>                    &lt;x&gt;-2051115&lt;/x&gt;<br>                    &lt;y&gt;2739663&lt;/y&gt;<br>                &lt;/point&gt;<br>            &lt;/pointlist&gt;<br>            &lt;specialtype&gt;rectangle&lt;/specialtype&gt;<br>        &lt;/annotation&gt;<br>    &lt;/ndpviewstate&gt;<br><br>I wonder if anyone dabbled in something like this and have an idea what sort of scaling +/- offsetting that they use for marking ROIs like this.<br><br>There is only one github module that may be of help in this conversion, but I failed to run the conversion code on linux and the project seems old/abandoned.<br><a href="https://github.com/Ichinaru/slideSelector">https://github.com/Ichinaru/slideSelector</a><br><br></div><div>Best wishes,<br>Ahmed<br><br></div></div>