Region Filling Robert Curro CIS 467-Image Processing
Region Filling Algorithm Binary Region Filling
Beginning with a point p inside of a boundary, fill the entire region with 1’s
GrayScale Region Filling
Beginning with a point p inside of a boundary, fill the entire region with the mean value of the region
This illustration uses the Matlab command:
imfill(im, ‘holes’)
Region Filling Definition Region filling is based on a set of dilations, complementation, and intersections. Xk = (Xk-1
B) ∩ Αc for k 1,2…n
Where (Xk-1 B) is a dilation B is the structuring element Αc is the complement of the image
Complementation Find the Complement of a binary image:
Original Image
Image Complement
Conditional Dilation Start From a point inside the region Dilate the point with the structuring element. Keep all points that intersect with the complement. Start
Structuring Element
Step by Step B
A
Ac
Start
X0
X1
X2
X6
X7
X7∪Α
Complementation, Dilation, Intersection
1. Complement the original image 2. Start at a location inside of a
boundary 3. Dilate the pixel with the structuring element 4. Keep all pixels that intersect with the complement of that image 5. Repeat from 3 until entire region is filled
Gray Scale Region Filling •Fills in Region with the mean value of that region
Problems Cannot determine background points from inner sphere points. Computer Vision algorithms need to be applied to find inner sphere points
Application:Find the Area of a Rat Lung
Application provided by Reindeer Graphics
Application: Segmentation of a Liver Threshold
Region Filling
Extract Largest Region
Extract Largest Region
Boundary Peeling
Application by Jie Zhu, Cornell University
Application: Finding an Outer Boundary of a Cell Find Edges
Original Image
Dilate Image
Region Filling
Application Continued…
Remove Connected Component on Boundary
Application by Mathworks, Matlab
Questions?