Modeling Respiratory Signals by Deformable Image Registration on 4DCT Lung Images

Algorithm 1

Remove artifacts.

1. Input the Matrix SliceIM (One lung slice image in each phase)

2. Find the Dark and Light Area in an image

   LightArea = find(image, ‘light’);

   DarkArea = find(image, ‘dark’);

3. Fill the DarkArea within a LightArea

   SliceIM = floodfill(SliceIM, LightArea, DarkArea);

4. Quantitate SliceIM into three indexed images using the Minimum Variance Quantization

   IndexIM = Quantization(SliceIM, 3)

5. Select the index partition in which it is the largest area

   MaxPartitionIM = MaxArea(IndexIM)

6. Fill the holes in MaxPartionIM to get the whole lung partition without artifacts

   OutputIM = FillHoles(MaxPartitionIM)

7. End

8. Result in the Matrix OutputIM (The lung slice image without artifacts, have only lung and body area)

Appendix

FillHoles method

1. Find the Dark and Light Area in an image

   LightArea = find(image, ‘light’);

   DarkArea = find(image, ‘dark’);

2. Fill the DarkArea within a LightArea by floodfill

   image = floodfill(image, LightArea, DarkArea);