Automatically selecting objects

If the mouse is double-clicked, tnimage tries to identify an object in the image. Since images are not in themselves composite objects, object identification is done by finding a group of contiguous pixels within a certain color range of the point at which the mouse was clicked. For example, if the mouse is double-clicked on a white area, tnimage will look for and select a white object centered at that point. If ``crawling dots'' are not displayed around the edges of the object after clicking on it, it means no object could be identified.

Once an object is selected, the selected portion can be copied, contrast-enhanced, etc. without affecting the surrounding pixels. For example, the white matter in the brain image below (from the Visible Human project) was changed from yellow to blue by simply double-clicking on the yellow area, and then inverting the colors of the selected area.

Changing color in non-rectangular portions of an image.

The ``Object threshold'' value (in the Config..Configure menu) determines how much surrounding area is included. Any pixels in the object whose red, green, and blue values differ from those at the starting point (i.e., the point on which you double-clicked) by less than the threshold are included. A lower value means the selection is more stringent. For example, increasing the red threshold and lowering the green and blue threshold will make it easier to select an object which is predominantly red.