Čeština (Česká republika)

ROI Analysis

The software load and analysis CT images from the DICOM format (Digital Imagin and Communications in Medicine). The program shown tissues in the Hounsfield units and is able to qualitatively compare selected areas from the CT images. The program is primarily design to determine bone density which is significant at creating a model material of the bone in biomechanics. Obtained data have got applications in a clinical practice..

The software was created as a part of the specific research projet at University of Technology in Brno FSI-J-10-70 and University Development Fund 1403/2010.

ROI Analysis HELP

The software is written in Matlab and to the run requires DLL libraries from this software. In this websites are two version, first written for Matlab 2008 (ROI Analysis08.exe) and second for Matlab 2009 (ROI Analysis09.exe). The program performs the CT images analysis of a selected ares - ROI (Region of Interest)

1. Description of the working area

The working area of the program is divided into four parts, fig. 1.1.



Fig 1.1 Description of the working area
1 - working area, 2 - information about images
3 - DICOM files loaded, 4 - Preference


2. Load images

Menu > File > Load CT (Ctrl + L)

The software load DICOM files with *.dcm suffix from a selected folder. Progress of the database loading can be viewed on the screen, Figure 2.1. After a successful database load, a message with a number of the loaded images appears on the screen.



Fig 2.1 Loading of images


3. Database save

Menu > File > Save as… (Ctrl + D)

The program save database of created defined areas (ROI) with all images. The save format is *.mat.


4. Database open

Menu > File > Open file… (Ctrl + O)

The program loads a previously saved database in *.mat without losing the previous defined ROI area.


5. Preferences

Menu > File > Preferences (Ctrl + N)


setting scale

limit bone in HU units for BVF (BV/TV) calculating

peak setting for 3D ROI graph

peak setting for 3D CT graph


6. Zoom tool and scroll tool

These tools are used to zoom and move with loaded image and has to be deactivated before define of ROI areas.


7. Switch between images

Between the CT images is possible to switch in the windows in fig. 7.1. Each image contains information important to idntify the patioent, used equipment and workplaces. Placing images on the PC is written bellow the main workspace.



Fig 7.1 Name and information of image


8. Color change

The program allows to change the black and white color to the pseudo color or to the warm colors which allows you to highlight different tissues in Fig 8.1.


  Fig 8.1 Image in various colors


9. Entering the ROI

If the ROI area is not defined, the highest point with value defined in preferences (Peak for 3D CT slices) can be displayed in 3D image. Likewise it can be displayed a 3D color image. With "define ROI" button, the user create the polygon which close by double-clicking or click on the first item. With created polygon can be moved (with the whole polygon or just with the points). The area is definitely confirms with another double-click, fig 9.1. Created areas is possible to delete: order making the ROI are marked by red numbers and the order in which they will delete by blue numbers.



Fig 9.1 Define and deleting of ROI


10. CT Analysis

The program allows to view all outputs at once after the check all boxes. For clarity is recommend gradually as necessary to paint each section separately. The graphs and the resulting of analysis will be open by Matlab package figure.


Fig 10.1 Options analysis of CT


The images can be stored in commonly used files (*.jpeg, *.png, etc.). For each CT image, the information belonging to it will plots and data from other CT will not draw.


10.1 3D with CT image

Show the 3D structure picture where the greatest value can be change in the preference (peak for 3D CT slices). The highest values correspond to the maximum depth of intensity pixels. The ROI areas are in the picture shown in yellow color.


Fig. 10.1 Rendered 3D CT image with marked areas


10.2 3D CT in color

Show the 3D structure picture in colors where the greatest value can be change in the preference (top of the 3D CT images). The ROI areas are diplayed in green.



Fig.10.2 3D CT image in the color


10.3 ROI in HU units

The program displays the selected ROI areas in hounsfield units according to which it is possible to determine the type of tissue and aualitatively assess the bone properities, fig. 10.3. The highest values of HU is +3000 (dental enamel) and the smallest is -1000 (air).


Obr 10.3 Hounsfield Units (HU)


Between the maximum and the minimum is allways a color scale. To visible differences between the transitions of HU units is changing the values only, fig 10.4.



Fig 10.4 Comparison of the three analysis of the ROI areas


10.4 3D ROI Analysis

For a clearer analysis is possible to show graphs in 3D. The highest value in the graph corresponds to the maximum value of intensity in the CT image (preferences - Peak for 3D graph ROI)



Fig. 10.5 ROI areas in 3D


10.5 Bone volume fraction BVF

This feature is primarily intended for the bone tissue analysis. The pgram determines dimensionless ratio between bone volume and tissue volume BV/TV (bone volume/tissue volume). It could be speaken about volume because each CT image has got thickness. An important criteria is in "Limit bone" setting (Chapter 5) which are expressed in units of HU value which becomes a pixel that corresponds to the bones. Limit deciding whether the pixel will or will not be counted to the bone tissue volume. The graph shows the progress BVF over the height for each row and total value is written above the graph, fig 10.6.


Fig 10.6 BVF – left „Limit bone“ 1000HU; right 1500HU


11 Sample of analysis

The HU values of CT images of a mandibula for cancellous bone in the length of 42.5 mm (teeth of mandibula) were analyzed with this software, fig 11.1. The HU units dependence on the length is shown on the graph fig 11.2. Three measurements were performed for the entire section of the cancellous bone and the alveolar and basal part of bone. The results are relevant for clinical practice to determine the implant.


Fig 11.1 Analyzed part of mandibula with ROI marked


Fig. 11.2 Behaviour of the HU units in the toothless mandila