Minute Mosaic Masterpiece - make a photomosaic using Flickr images!

All instructions for installing on Linux systems; apologies to Windows users. Python 2.7

Pip: sudo apt-get install pip

libjpeg: sudo apt-get install libjpeg8 libjpeg8-dev

pygame: sudo apt-get install python-pygame

ujson: sudo pip install ujson

flickrapi: sudo pip install flickrapi

elementtree: sudo pip install elementtree

numpy: sudo pip install numpy

stemming: sudo pip install stemming

For crawling Flickr to obtain pictures. Must replace APIKEY and APISECRET with your own key and secret.

RUN: python khittykrawler.py

This will take a long time and lots of storage space.

Requires files BatchProcess.cpp, ImageAnalysis.cpp, ImageAnalysis.h

COMPILE: g++ -o batch BatchProcess.cpp ImageAnalysis.cpp -ljpeg -fpermissive

RUN: ./batch

Pictures to analyze should be in ./pictures; output will be placed in ./pictures/processed. You must create processed before running ./batch.

Output is colordata.json, cropped/resized images in ./pictures/processed.

Demonstrates the algorithm for determining the closest color to query color

You must have pygame installed to view this demo

RUN: python scoring.py colordata.json

You will see a textbox looking graphic along with rows of circles of colors given in the data.

Using the numbers about QWERTY (keypad will not register), enter a RGB value in this format: 123,123,123 Include the commas and enter a correct RGB value or the program will crash. You can use the backspace to clear the box. Then press Enter

You will see a circle of the color you entered appear to the right of the textbox and the closest color in the data will be have a white circle drawn around it.

You can continue to query different colors by deleting your previous RGB value, entering another, and hitting Enter again.

Press ESC to exit the application

Analyzes the tags of all the pictures and arranges them into corresponding hierarchical clusters

You must have numpy installed These file must also be included: tfidf.py cluster3.py cluster_tests.py datasample.json

RUN: nosetests.py -s

This test file consists of 5 different tests. The first two make sure that similar values are grouped into the same cluster. the third test shows that the algorithm starts out with every picture being in its own cluster the fourth test demonstrates that at the end of the algorithm there are only two main clusters. the last test proves that every picture is included in a cluster