Computer Science 15-111 (Sections A & B), Spring 2007

Homework #4.1

Due: Fri 02-Feb-2007 at 8:00pm (online submission).

There is no written portion to this assignment, just the following code which should be place in hw4.1.zip.

1. Image Editor
Note that our PixelImages.java file has been modified to load a pixel array from a jpeg file (see the sample drawing which displays this sampleImage.jpg, and the comment at the head of the file, for more details). Starting from this version of PixelImages.java, modify it by adding the following methods:

public static int[][] rotateImage(int[][] pixels, int degrees)
This method rotates the image counterclockwise by the given amount, where “degrees” must be a multiple of 90 (if not, you should round it to the nearest multiple of 90, whether positive or negative). Note that rotating a non-square image will change its dimensions (why?), which is not possible to do with an int[][] parameter (again, why?), so this method does not modify the pixels array! Instead, it returns a new pixel array. Thus, it is used as such:
// rotated 90 degrees counterclockwise
int[][] newPixels = rotateImage(pixels,90);

public static int[][] cropImage(int[][] pixels,
int left, int top,
int width, int height)
This method does not modify the pixels array itself, but returns a new pixel array of dimensions (width x height) containing just those pixels in the given rectangle of the source image.

public static int[][] zoomInImage(int[][] pixels, int zoom)
This method does not modify the pixels array itself, but returns a new pixel array of dimensions (zoom * width, zoom * height), where width and height refer to the dimensions of the source array, and where “zoom” is a positive integer (you do not have to handle the negative or zero case). To enlarge an image, each source pixel will be mapped to a zoom-by-zoom block of pixels in the target. So, if zoom==3, each source pixel would map to a 3x3 block of the same pixel in the target. Note that this may look a little “pixelated”, and we could improve this by smoothing the transitions between pixel blocks, but that is not necessary for this assignment.

public static int[][] zoomOutImage(int[][] pixels, int zoom)
This method (which essentially makes thumbnails, as we did in class!) does not modify the pixels array itself, but returns a new pixel array of dimensions (width / zoom, height / zoom), where width and height refer to the dimensions of the source array, and where “zoom” is a positive integer (you do not have to handle the negative or zero case). Shrinking an image is the opposite of enlarging it. Here, each zoom-by-zoom block of source pixels is mapped to a single target pixel. To do this, take the average value of the alpha, red, green, and blue over the entire source block, and use these values for the target pixel. Note that you must average by individual ARGB component, and not by the entire “int” value!

public static void testHw4()
This method should read in the sampleImage.jpg into a pixels array, then display 12 images as follows:
a) Display the unmodified jpeg;
b) Rotate the jpeg by 90, 180, and 270 degrees, displaying each result;
c) Crop the image to only display the middle portion (those pixels not in the topmost or bottommost quarters when divided vertically, nor in the leftmost or rightmost quarters when divided horizontally), and display this result;
d) Zoom in by 2 times, 3 times, and 4 times, and display each result;
e) Zoom out by 2 times, 3 times, and 4 times, and display each result;
f) Redisplay the original jpeg, demonstrating that it was unmodified throughout this method.