CSC352 Danaë's Project 2013

From CSclasswiki
Jump to: navigation, search

Here's a basic overview of what I'm doing...

Input: 3 million "images" (just plain blocks), with random ID numbers, names, frequencies, etc. Output: An efficient and aesthetic packing of those images.

Steps:

1. pre-process

- fill a mySQL database with the images and their information

2. process

- partition images into 9 groups based on their frequencies (not a uniform distribution; there should be a large number of less important images and a small number of more important images).

- scale the images according to which partition each is in (not according to its actual importance, but according to the group it was placed into)

3. pack

- pre-pack the most important partition

- use threads (trying different numbers of threads for efficiency) to pack the rest of the images, each thread working within one of nine partitions of the final product.