ImageP 0.292

Image processing add-ons to Python and Numpy
With ImageP you can do scientific image analysis with Python/iPython, Numpy and Matplotlib easily, providing an open source, free alternative to IDL and Matlab image processing.

Developer comments

This started out as a collection of short functions I needed for my image processing purposes. There are many possible libs out there, some of them really powerful. I wanted something simpler, easy to install and tweaked one.

Some corner points:
- minimize dependencies to numpy, matplotlib (and in the worst case scipy)
- keep it as small and clear as possible - scientific work requires transparency
- the core is a numpy array: do everything on this without any extra conversions
- write all in python, some speedups can be added in C keeping the python equivalent (useful for learning and understanding)
- the C part at the moment is a simple dynamic lib communicating through ctypes and using numpy arrays directly (quick and dirty + trouble on Windows)
- I myself am learning through this process, I hope it helps others as well

last updated on:
August 3rd, 2010, 13:19 GMT
price:
FREE!
developed by:
Tamas Haraszti
license type:
GPL v3 
category:
ROOT \ Science and Engineering \ Image Recognition

FREE!

In a hurry? Add it to your Download Basket!

user rating

UNRATED
0.0/5
 

0/5

What's New in This Release:
  • SimpleFilter is a pixel-by-pixel convolution for small kernels (< 10 pixels possibly)
  • 1D version of SimpleFilter and ConvFilter for kernels which can be built by applying a 1D kernel in two directions separately. For these kernels (GaussKernel, BinomKernel etc) this is much faster.
  • GaussKernel, BinomKernel, BoxcarKernel can also generate 1D versions of themselves.
  • The norm=True is the default for the kernels now.
read full changelog

Add your review!

SUBMIT