In this contest we provide two tools. The first to generate test cases from images, and the second to visualize your solution.

Files

Blur.class -- test generator
Tester.jar -- tester
examples.tar -- examples
contact.html -- thumbnails of more tests
examples.tgz -- NEW updated examples

Test Generator

This program reads an image and produces a number of low resolution, blurred copies of that image. It places the copies in the same directory as the original, with "_0", "_1", ... appended to the filename before the prefix. To use it, simply copy Blur.class to your machine, and run the following in the directory you downloaded it to. 
    java -Xmx512M Blur <file> -D <D> -N <N> -SD <SD>
The -D and -N parameters are optional and will be randomly generated based on the filename if you omit them. Note that this can be very slow for large images.

Test Runner

As in past contests, you must write a program that communicates its input and output via stdin and stdout. The parameter N, along with the input parameters (W, H, retW, retH, D, and input) will be given to you on standard in, delimited by spaces, with N first followed by the other parameters in order. The int array, input will simply be given one element at a time (you know its length as W*H*N). Once you have read your parameters, your should simply output retW*retH integers, separated by whitespace (make sure you flush stdout). Anything you print to standard error will show up in the console.

To run the tool, you should run: 
    java -Xmx512M -jar Tester.jar
There are a number of optional parameters: When you start up the program will try to read "example.png" by default (unless you specify a different file). It will also try to read "example_0.png", "example_1.png", .... When you run your program, it will pass the parameters and wait for your return. Once it has received your image, it will attempt to align your image with the original. Once this is done, your score (relative to the baseline) will appear in the console. You may toggle your image on and off with the checkbox. You may also view the difference between your image and the original once the alignment is finished. You can click and drag to pan the image, and use the mouse wheel to zoom also.

The parameter for a test are all determined from the sizes of the images. The only one which you can change is the D parameter. If you didn't specify D while generating images, or you are testing the examples provided, you need not specify D in the Tester tool. You only need to specify it if you generated your own images with your own value of D.

Final Notes