Image Class Design and Implementation

The assignment is to create a C++ class for images that hides the details of the image data structure, allowing other code to use the class without changes when the underlying data structure is modified. The class should include constructors, copy constructor, destructor, operator=, accessor and mutator functions, output method, comparison operators, and methods for creating mirror images and photonegatives. A sample program demonstrates how to read an image, output its dimensions, create a mirror-image, modify the image, write it back to file, compare the original and modified images, and count differences if they are not equal.

Classes and object-oriented programming
If I changed the image data structure used in ImageLib.h and ImageLib.lib, all of the code that uses these
files would need to be changed. Your goal for this program is to create a C++ class for images that hides
the details of the image data structure. If the data structure is changed, only the details of your image class
should need to be changed (and not the interface), so that other code that uses the image class can be used
unchanged. This will also upgrade images to first-class objects with a constructor, destructor, operator=,
etc. Make sure that you use the concepts of data abstraction, information hiding, modularity, and fail-safe
programming whenever possible.
Assignment
You are to create a new class representing images. Remember that your project should be created as an
“Empty project,” so that you don’t get precompiled header code among other things. In order to create this
class you should continue to use the files for working with images from the previous programming
assignment:
http://courses.washington.edu/css501/cfolson/program1/ImageLib.h
http://courses.washington.edu/css501/cfolson/program1/ImageLib.lib
http://courses.washington.edu/css501/cfolson/program1/ImageLib.pdb
If you want a new test image, you can use this one:
http://courses.washington.edu/css501/cfolson/program2/test.gif
The assignment has two parts. First, you will play the role of class designer and create a new class for
images. Second, you will play the role of developer using the class library. For the second part, you should
use your class, but not make any assumptions about the implementation of the class (and don’t write new
member operations to solve that part).
Your image class should include (at least) the following capabilities:
 Constructor using a filename (read from a file)
 Constructor using just the number of rows and columns (create a blank, black image)
 Copy constructor, destructor, operator= (these are very important!)

 Accessor and mutator functions for individual image pixels
 Output to a file
 Overload the == and != operator for comparison of images (every pixel must be the same)
 Overload the << operator to output the number of rows and columns in the image (ignore the
pixels)
 Create a method that returns an image with the same dimensions as the original image, but as a
mirror image. It should be flipped horizontally, as you did in the first program. (The return value
should be an object of your new class returned by value. Do not modify the input image.)
 Create a method that returns the photonegative of the input. (Once again, do not change the input.)
Use your image class to write a program that performs the following operations using your new methods.
1. Read the image “test.gif”.
2. Output the number of rows and columns using operator<<.
3. Create a mirror-image using your member function (use this image as the input to step 4).
4. For every pixel, add col mod 5 to the red component and subtract row mod 7 from the green
component. Continue to check for (and correct) overflow and underflow.
5. Write the image as “output.gif” and read that file back into a new variable.
6. Compare the image you wrote to the image read in the previous step using your == operator for
images.
7. If the images are different, count the number of different pixels.
8. Output the number of differences.
Note that steps 4 and 7 should be performed using the basic class member functions you have already
written, do not write new member operations that perform these operations.
You generally won’t want to pass an image parameter by value, since this results in copying the entire
image. Passing by reference (using const when possible) is better.
For additional fun (optional)
Implement additional methods that operate on images. Some that you might try:
 For each pixel, replace it with the average color of the neighboring pixels.
 Stretch (or shrink) the image by some arbitrary scale factor.
 Compute the average of two images (pixel-by-pixel).
 Combine two images by creating a new image with the input images side-by-side.
 Create a new image that changes every color value to 0 or 255 depending on which it is closest to.
 Write an algorithm to determine whether there is a cat in the image.