We have developed a font-end tool - image object labeler, which enables
us to segment a forest image into its various constituents so that they
can be distinctly labeled. The manual transcription work generates
training and reference data for the automatic image segmentation task.

<br><br>

Here is the Tcl-Tk wrapper interface of the labeler (click for magnification):
<br><br>

<a href="../images/labeler_fullsize.gif">
<img src="../images/labeler.gif" alt="screenshot" border=0
align=center hspace=20></a>

<br><br>

Some example segmentations:
<br><br>

       <table align=center border=0 bgcolor="#FFFFFF" cellpadding=5>

       <tr align=center>
       <td>
         <img src="../images/polygon_seg_01.gif" alt="Example 1">
       </td>
       <td>
         <img src="../images/polygon_seg_02.gif" alt="Example 2">
       </td>
       </tr>

       </table>