Learning classification object selection issue

rtaliano

Usually after creating classes in the learning classification and then selecting a class for which one wants to pick the objects for learning, the program highlights the object being selected by four squares, one at each corner of the object's bounding box/rectangle. Under these conditions I can select as many objects as I want by holding down the control key. But periodically, for no reason I can identify, when I select objects it highlights the chose object by placing evenly spaced white boxes around the contour of the object. Under these conditions it is only possible to select one object at a time. Even when using the control key, the program highlights only the most recent object selected. The previously selected object is deselected by the program. The only way I have solved this issue before is to restart the program. But my current project has thousands of objects which makes the counting and splitting process very time consuming. Thus this is not a good option. Is some setting that can be changed to solve this problem?

Thanks
Ross Taliano

YuriG

Hi Ross,

The identification of selected objects, whether a rectangle or polygon, depends on the outline type, so for Filled it's a bitmap, for not Filled - a polygon.
If you have more that 10000 objects on the image, the overlay is shown as a single bitmap (to speedup drawings), so objects cannot be selected individually. 
If you have more that 10000 objects on the image and want to run learning classification, you should increase "Individual Object Display limit" in the application options (Application page).

Yuri

CraigW

When a polygon is the selected object, its vertices are marked with little selection squares.  You should be able to select multiple polygons with a Ctrl-Click in the interior of the polygon (you might have to zoom in to see the interior if the object is very small).  Also multiple objects can be selected by clicking and dragging across a number of objects (this will select all objects within the rectangular bounds of the selection).