mango314 / colores Goto Github PK

Just thought we should document how to get up and running on Mac OS X. After setting up my virtualenv I had to install via pip the following:

• numpy
• scipy
• scikit-image
• cython
• matplotlib
• Pillow (instead of PIL for better image format support on matplotlib)

I also had to run brew install gcc to get some fortran requirements installed for one of the above requirements.

Wouldn't it be great to return color-quantized versions of the original photograph. Be just like a paint-by-numbers :-)

How can we have this read an image and return the X number most common colors found?

By adding plt.savefig('filename-fig.jpg') at the end of colores.py I was able to export this file by running $ python colores.py

I imagine someone would use this as a package, which seems way more useful:

from colores import Colores

colores = Colores('filename.jpg')
list_of_hex_values = colores.extract_colors(max=5)
colores.save_figure('filename-fig.jpg') # Save a figure from pyplot.

If we upload the entire photograph get memory error. Optimize so reasonable photographs (up to 6MB) can be processed quickly.

Something that can be installed using pip