1) Objects detection and quantitative analysis

A "classic" example of blobs segmentation by thresholding. Pixels exceeding given threshold inside the user-drawn region of interest create connected regions. Contours of such regions are detected, converted into the polygonal representation and painted as "overlay". Detected objects can be combined with user drawn ones (line, angle, multiline, rectangle, polygon , freehand drawn polygon). Each object (or its part) can be manually edited (moved, deleted, reshaped etc.). Right picture on figure bellow represents zoomed version of left image, where one point of object No.2 outline was moved into another position.

Quantitative analysis of detected objects is performed when the Results Window is opened (bottom window). More than 15 characteristic parameters is calculated and displayed in Results Window in calibrated units (area, perimeter, Roundness, Ferret diameter, density values, center of gravity, moments etc.). There is direct interaction between the active object on image and an active row in the table (both are painted by the same color - blue in our example). The change of active object shape is immediately followed by the recalculation of values in Results Window. Selected cells of Results Window (yellow ones) can be transferred into another programs via clipboard (the whole table via text file).

2) Nano-particles analysis

In some special applications (e.g. magnetic nano-particles studied by Atomic Force Microscope), the particle size is calculated as the geometrical mean of sizes in X,Y,Z directions. The particle "height" is obtained as the brightness difference between the background and the top of the particle  (recalculated according to the density calibration curve). The XY sizes correspond to the Ferret diameter measured in one half of the height. Starting from the threshold-based detection of outlines (left image), the algorithm automatically searches for two outlines shown on the right image - the first one lies just above the background (white) and the second one is detected in one half of height (black). This task is performed by a special AFMParticle module developed at the Institute of Experimental Physics, Slovak Academy of Sciences, Kosice.

3) Serial sections alignment

An obvious problem of serial sections analysis in microscopy is registration (alignment of two images, rotated and translated each-other). Automated algorithms based on the correlation analysis usually fail in case of low-quality images and therefore a simple tool for semi-interactive alignment is useful. Module Align allows both coarse and fine alignment of a stack of images.  The thresholded red image is compared with the neighbor in stack (blue one), the overlapping parts are displayed black. The operator can search for the best match of images either by re-enter of calibration marks or by using of navigation buttons.

4)  Length measurement of DNA plasmids

Atomic force microscope (AFM) allows study of individual DNA chains. The special program module DNATracker allows semi-automatic tracking of individual DNA chains from the initial position pointed by the operator. In well-defined parts, algorithm follows the "ridge" representing DNA, in critical parts (crossing, touching and overlapping structures) operator navigates the algorithm pointing the direction. Images capturing and analysis was done at the Institute of Condensed Matter Physics, Lausane University, Switzerland. DNATracker plug-in module based on Ellipse environment has been developed by the Institute of Experimental Physics, SAS, Kosice.

5) Modules for the stereological analysis

Stereology estimates 3D particles from their 2D sections. The most of stereological plug-in modules based on Ellipse environment have been developed cooperating with Dept. of Biomathematics, Institute of Physiology, Academy of Sciences of the Czech Republic. Contact the authors to get information about individual stereological modules (Fakir, Slicer, PointGrid, Disector, ClosedSpline).

The figure bellow shows an example of PointGrid module function. If the cross representing grid element hits an objects, it is marked as foreground (green), otherwise it is background (yellow). The initial setting is based on the thresholding of pixel value "under" the cross. However, real images are usually much worse than popular "blobs" image below and therefore the manual correction of thresholded values is necessary (set / reset point by clicking it, or by outlining some area by the rectangle or by the polygon).

6) Visualization and analysis in 3D

Starting from version 2.05, Ellipse can be supplied with optional 3D modules. This extended version offers a new quality in 3D and therefore a new name "Ellipse3D" was used.

Ellipse3D allows:

• visualization and rendering of object segmented in stack
  
• rotation and zoom of reconstructed object in 3D (see figure bellow where a few selected positions are displayed)
  
• change of a material properties (transparency, color, etc...)