README

phash-graph-mvp is a small python script which use pHash ( http://www.phash.org/ )
to build a graphical representation of perceptual relation between images.

Say you have a directory full of similar (cropped, blur, noise, filters) images.
You can create a graphical representation of relation between pictures.

Example:

 $ python graphmvp.py 
ERROR:root:usage1: graphmvp.py <dirname> <dbname> <outputgraph> # build graph from database
ERROR:root:usage2: graphmvp.py <dirname>                        # build database and graph (beta)

Build a database and create a graph
 $ python graphmvp.py examples/datasets/

Outputs :
datasets.db.mvp   -> pHash MVP Database.
datasets.db1.mvp  -> pHash MVP Database.
datasets.png.dot  -> pygraphiz graph representation file
datasets.png      -> PNG output



Create a graph from a database and a directory.
 $ python graphmvp.py examples/datasets/ datasets.db datasets.png

Inputs:
datasets.db.mvp   -> pHash MVP Database.
datasets.db1.mvp  -> pHash MVP Database.
examples/datasets -> the source folder with the image used in building the MVP database 

Outputs :
datasets.png.dot  -> pygraphiz graph representation file
datasets.png      -> PNG output