26. CT: IMAGE RECONSTRUCTION

Question 1

1. What are the 8 steps to reconstructing images from an X-Ray Measurement?

NB:
- this is known as the Convolution Back Projection
Procedure

Answer 1

1. Measurements
2. Pre-Processing
3. Raw Data
4. Filtering
5. Filtered Data
6. Interpolation
7. Back-Projection
8. Axial Source Images

Question 2

2. What happen when we have selected a “Field of View”?

Answer 2

• it is divided into small image elements
• it is 2D
• these are called PIXELS

Question 3

3. What do the Pixels that make up each cross-sectional image represent?

Answer 3

• they represent a small volume of tissue
• this is called a Voxel
• this stands for Volume Element
• we take into consideration the depth of the Pixel
• this takes into consideration the volume of the Pixel

Question 4

4. What can be said about a Tomographic Axial image that consists of a matrix of:
   512 x 512 x 1024 x 1024 Voxels ?

Answer 4

• this means that over between 260 000 - 1 million
  grayscale values must be stored
• they are stored for each cross-sectional image

Question 5

5. In which image are the Pixels more evident?

Answer 5

THE IMAGE TO THE LEFT:
- this is an older image
- it has a low-resolution

THE IMAGE TO THE RIGHT:
- has pixels that are much less evident
- it is a more modern image
- it has a higher resolution

Question 6

6. What does the density of each Pixel depend on?

Answer 6

• it depends on the Composition of the tissue it
  represents
• it is expressed in a Hounsfield Unit (HU)

Question 7

7. How are Hounsfield units calculated?

Answer 7

• they are calculated from the Attenuation
  Measurements that are relative to the Attenuation of
  water

NB:
- the attenuation of water is Zero

Question 8

8. What values do the Hounsfield units range from?

Answer 8

THEY RANGE FROM:
- negative 1024 to + 3071

NEGATIVE VALUE:
- no absorption

HIGH POSITIVE VALUES:
- high absorption

Question 9

9. Does this diagram make sense?

NB:
- know this scale very well

Answer 9

• yes

NB:
- know this scale very well

Question 10

10. Does this diagram make sense?

Answer 10

• yes

Question 11

11. What is the main factor of the Convolution Back Projection Procedure?

Answer 11

• we obtain the Raw Data from the measurements
• we filter the Raw Data to remove the image noise
• this obtains the filtered data
• the contrast and spacial resolution are increased
• a better quality image is obtained

Question 12

12. What forms the image matrix?

Answer 12

• the pixels
• the voxels

Question 13

13. What does the Image Matrix affect?

Answer 13

THE RESOLUTION:
- increases when there are more pixels and voxels
present

THE TIME:
- it takes more time to construct an image when there
are more pixels and voxels present

Question 14

14. What is Lambert Beer’s Law?

Answer 14

• there is a linear relationship between the
  concentration and the absorbance of the solution
• this linear relationship can be calculated by measuring
  the absorbance