Connected-Component Labeling and Set Operations

As discussed in L5, this project considers the application of a sequence of simple image-processing operations to an image, such as connected-component labeling and logical (set) operations.

1. Bright-Region Extraction
   • Consider the gray-scale image, lake.gif. Experimentally choose a threshold so that > 4 distinct bright components obviously appear. Save this thresholded image, fthresh, in a form that makes the thresholded objects visible on the screen.
   • Find the connected components of the thresholded image fthresh. You can use the MATLAB function bwlabel for this purpose. The following MATLAB call creates a labeled image called flabel from the input thresholded (binary valued!) image fthresh:

[flabel, num] = bwlabel(fthresh, 8) where 8-connectivity is assumed for the components and num is the number of connected components labeled in fthresh. To display this labeled image with colored components, you can use fRGB = label2rgb(flabel);

And then use imshow(fRGB) to see the colored labeled image.

• Save the 2 largest components of your labeled image and delete the other components by setting their constituent pixels to 0. You will need to write a function to do this.
• Be sure to give output images for all steps above.

2. Logical (Set) Operations

Note: you are to write your own functions for the operations in this part of the project you may NOT

use built-in Matlab functions.

• Write Matlab functions for the AND, OR and XOR binary-image operators and NOT unary-image operator, using A and B as input images. What are the quantities A AND B , A OR B, A XOR B, and NOT(A) in terms of set union, intersection, and complement?
• Let A be the match1 image and B be the match2 image. Compute the following images: A AND

B , A OR B, A XOR B, and NOT(A)

• Build the minimum operator and compute E = min(C,D), where image C is mandrill gray and image D is cameraman. For each pair of pixels (x,y) in the two input images, the minimum operator assigns the minimum of the two values (C(x,y) and D(x,y)) to the output E(x,y). As we will see later during our discussion of Morphological Image Processing (G&W Ch. 9), the minimum operator is a gray-scale analog of set intersection (AND) and is sometimes called erosion.

3. Write a report in the standard format. Be sure to describe all of your methods, including 1(c) and other parts.