skip to navigation
skip to content

mahotas 0.7

Mahotas: Computer Vision Library

Package Documentation

Latest Version: 1.4.4

Image Processing Library for Python.

It includes a couple of algorithms implemented in C++ for speed while operating
in numpy arrays.

Notable algorithms:
- watershed.
- convex points calculations.
- hit & miss. thinning.
- Zernike & Haralick, LBP, and TAS features.
- freeimage based numpy image loading (requires freeimage libraries to be
- Speeded-Up Robust Features (SURF), a form of local features.
- thresholding.
- convolution.
- Sobel edge detection.


This is a simple example of loading a file (called `test.jpeg`) and calling
`watershed` using above threshold regions as a seed (we use Otsu to define


import numpy as np
from scipy import ndimage
import mahotas
import pylab

img = mahotas.imread('test.jpeg')
T_otsu = mahotas.thresholding.otsu(img)
seeds,_ = ndimage.label(img > T_otsu)
labeled = mahotas.cwatershed(img.max() - img, seeds)


Recent Changes

For version **0.7 (Dec 5 2011)**:

The big change was that the *dependency on scipy was removed*. As part of this
process, the interpolate submodule was added. A few important bug fixes as

- Allow specification of centre in Zernike moment computation
- Fix Local Binary Patterns
- Remove dependency on scipy
- Add interpolate module (from scipy.ndimage)
- Add labeled_sum & labeled_sizes
- gvoronoi no longer depends on scipy
- mahotas is importable without scipy
- Fix bugs in 2D TAS (reported by Jenn Bakal)
- Support for 1-bit monochrome image loading with freeimage
- Fix GIL handling on errors (reported by Gareth McCaughan)
- Fix freeimage for 64-bit computers

For version **0.6.6 (August 8 2011)**:
- Fix fill_polygon bug (fix by joferkington)
- Fix Haralick feature 6 (fix by Rita Simões)
- Implement ``morph.get_structuring_element`` for ndim > 2. This implies that
functions such as ``label()`` now also work in multiple dimensions
- Add median filter & ``rank_filter`` functions
- Add template_match function
- Refactor by use of mahotas.internal
- Better error message for when the compiled modules cannot be loaded
- Update contact email. All docs in numpydoc format now.

For version **0.6.5**:
- Add ``max_points`` & ``descriptor_only`` arguments to
- Fix haralick for 3-D images (bug report by Rita Simões)
- Better error messages
- Fix hit&miss for non-boolean inputs
- Add ``label()`` function

For version **0.6.4**:

- Fix bug in ``cwatershed()`` when using return_lines=1
- Fix bug in ``cwatershed()`` when using equivalent types for image and markers
- Move tests to mahotas.tests and include them in distribution
- Include ChangeLog in distribution
- Fix compilation on the Mac OS
- Fix compilation warnings on gcc

For version **0.6.3**:

- Improve ``mahotas.stretch()`` function
- Fix corner case in surf (when determinant was zero)
- ``threshold`` argument in
- imreadfromblob() & imsavetoblob() functions
- ``max_points`` argument for
- Add ``mahotas.labeled.borders`` function

For version **0.6.2**:

Bugfix release:

- Fix memory leak in _surf
- More robust searching for freeimage
- More functions in to retrieve intermediate results
- Improve compilation on Windows (patches by Christoph Gohlke)

For version **0.6.1**:

- Release the GIL in morphological functions
- Convolution
- just_filter option in edge.sobel()
- mahotas.labeled functions
- SURF local features

For version **0.6**:

- Improve Local Binary patterns (faster and better interface)
- Much faster erode() (10x faster)
- Faster dilate() (2x faster)
- TAS for 3D images
- Haralick for 3D images


*Website*: `
<http:"" software="" mahotas="">`_

*API Docs*: `
<http:"" mahotas=""/>`_

*Mailing List*: Use the `pythonvision mailing list
<http:"" group="" pythonvision?pli="1">`_ for questions, bug
submissions, etc.  
File Type Py Version Uploaded on Size
mahotas-0.7.tar.gz (md5) Source 2011-12-06 95KB