wassname/scikit-image
1
0
Fork 0
mirror of https://github.com/wassname/scikit-image.git synced 2026-07-07 10:24:25 +08:00
Code Issues Packages Projects Releases Wiki Activity

Minor code cleanup.

Browse Source
This commit is contained in:
Stefan van der Walt
2012-12-11 18:14:45 -08:00
parent 530fbbb2fc
commit 7ebbdfd75f
1 changed files with 72 additions and 75 deletions
Download Patch File Download Diff File Expand all files Collapse all files
skimage/exposure/_adapthist.py
+72 -75
View File
@@ -1,26 +1,18 @@
'''
Adapted code from the article
* "Contrast Limited Adaptive Histogram Equalization"
* by Karel Zuiderveld, karel@cv.ruu.nl
* in "Graphics Gems IV", Academic Press, 1994
=============
"""
Adapted code from "Contrast Limited Adaptive Histogram Equalization" by Karel
Zuiderveld <karel@cv.ruu.nl>, Graphics Gems IV, Academic Press, 1994.
http://tog.acm.org/resources/GraphicsGems/
http://tog.acm.org/resources/GraphicsGems/gems.html#gemsvi
EULA: The Graphics Gems code is copyright-protected.
In other words, you cannot claim the text of the code as your
own and resell it. Using the code is permitted in any program,
product, or library, non-commercial or commercial.
Giving credit is not required, though is a nice gesture.
The code comes as-is, and if there are any flaws or problems
with any Gems code, nobody involved with Gems - authors, editors,
publishers, or webmasters - are to be held responsible.
Basically, don't be a jerk, and remember that anything free
The Graphics Gems code is copyright-protected. In other words, you cannot
claim the text of the code as your own and resell it. Using the code is
permitted in any program, product, or library, non-commercial or commercial.
Giving credit is not required, though is a nice gesture. The code comes as-is,
and if there are any flaws or problems with any Gems code, nobody involved with
Gems - authors, editors, publishers, or webmasters - are to be held
responsible. Basically, don't be a jerk, and remember that anything free
comes with no guarantee.
*
* Author: Karel Zuiderveld, Computer Vision Research Group,
* Utrecht, The Netherlands (karel@cv.ruu.nl)
'''
"""
import numpy as np
import skimage
from skimage import color
@@ -30,41 +22,37 @@ from skimage.util import view_as_blocks
MAX_REG_X = 16 # max. # contextual regions in x-direction */
MAX_REG_Y = 16 # max. # contextual regions in y-direction */
NR_OF_GREY = 1 << 14 # number of grayscale levels to use in CLAHE algorithm
NR_OF_GREY = 16384 # number of grayscale levels to use in CLAHE algorithm
def equalize_adapthist(image, ntiles_x=8, ntiles_y=8, clip_limit=0.01,
nbins=256):
'''Contrast Limited Adaptive Histogram Equalization
"""Contrast Limited Adaptive Histogram Equalization.
Parameters
----------
image : array-like
original image
Input image.
ntiles_x : int, optional
Tile regions in the X direction (2, 16)
Number of tile regions in the X direction. Ranges between 2 and 16.
ntiles_y : int, optional
Tile regions in the Y direction (2, 16)
Number of tile regions in the Y direction. Ranges between 2 and 16.
clip_limit : float: optional
Normalized cliplimit (higher values give more contrast)
Clipping limit, normalized between 0 and 1 (higher values give more
contrast).
nbins : int, optional
Greybins for histogram ("dynamic range")
Number of gray bins for histogram ("dynamic range").
Returns
-------
out : np.ndarray
equalized image - grayscale images are uint16, color images are float
out : ndarray
Equalized image.
Notes
-----
* The algorithm relies on an image whose rows and columns are even
multiples of the number of tiles, so the extra rows and columns are left
at their original values, thus preserving the input image shape.
<<<<<<< HEAD
=======
* For grayscale images, CLAHE is performed on one channel,
and a grayscale is returned
>>>>>>> 2e1729a9fbbc21fc0b04df8e68efbab9cfd6dada
* For color images, the following steps are performed:
- The image is converted to LAB color space
- The CLAHE algorithm is run on the L channel
@@ -73,10 +61,9 @@ def equalize_adapthist(image, ntiles_x=8, ntiles_y=8, clip_limit=0.01,
References
----------
.. [1] http://tog.acm.org/resources/GraphicsGems/
.. [1] http://tog.acm.org/resources/GraphicsGems/gems.html#gemsvi
.. [2] https://en.wikipedia.org/wiki/CLAHE#CLAHE
'''
# handle color images - CLAHE accepts scalar images only
"""
args = [None, ntiles_x, ntiles_y, clip_limit * nbins, nbins]
if image.ndim > 2:
lab_img = color.rgb2lab(skimage.img_as_float(image))
@@ -99,31 +86,32 @@ def equalize_adapthist(image, ntiles_x=8, ntiles_y=8, clip_limit=0.01,
def _clahe(image, ntiles_x, ntiles_y, clip_limit, nbins=128):
'''Contrast Limited Adaptive Histogram Equalization
"""Contrast Limited Adaptive Histogram Equalization.
Parameters
----------
image : array-like
original image
Input image.
ntiles_x : int, optional
Tile regions in the X direction (2, 16)
Number of tile regions in the X direction. Ranges between 2 and 16.
ntiles_y : int, optional
Tile regions in the Y direction (2, 16)
clip_limit : float: optional
Normalized cliplimit (higher values give more contrast)
Number of tile regions in the Y direction. Ranges between 2 and 16.
clip_limit : float, optional
Normalized clipping limit (higher values give more contrast).
nbins : int, optional
Greybins for histogram ("dynamic range")
Number of gray bins for histogram ("dynamic range").
Returns
-------
out : np.ndarray
out : ndarray
Equalized image.
The number of "effective" greylevels in the output image is set by nbins;
The number of "effective" greylevels in the output image is set by `nbins`;
selecting a small value (eg. 128) speeds up processing and still produce
an output image of good quality. The output image will have the same
minimum and maximum value as the input image. A clip limit smaller than 1
results in standard (non-contrast limited) AHE.
'''
"""
ntiles_x = min(ntiles_x, MAX_REG_X)
ntiles_y = min(ntiles_y, MAX_REG_Y)
ntiles_y = max(ntiles_x, 2)
@@ -148,10 +136,12 @@ def _clahe(image, ntiles_x, ntiles_y, clip_limit, nbins=128):
clip_limit = 1
else:
clip_limit = NR_OF_GREY # Large value, do not clip (AHE)
bin_size = 1 + NR_OF_GREY / nbins
aLUT = np.arange(NR_OF_GREY)
aLUT /= bin_size
img_blocks = view_as_blocks(image, (y_size, x_size))
# Calculate greylevel mappings for each contextual region
for y in range(ntiles_y):
for x in range(ntiles_x):
@@ -162,6 +152,7 @@ def _clahe(image, ntiles_x, ntiles_y, clip_limit, nbins=128):
hist = clip_histogram(hist, clip_limit)
hist = map_histogram(hist, 0, NR_OF_GREY, n_pixels)
map_array[y, x] = hist
# Interpolate greylevel mappings to get CLAHE image
ystart = 0
for y in range(ntiles_y + 1):
@@ -178,6 +169,7 @@ def _clahe(image, ntiles_x, ntiles_y, clip_limit, nbins=128):
ystep = y_size
yU = y - 1
yB = yB + 1
for x in range(ntiles_x + 1):
if x == 0: # special case: left column
xstep = x_size / 2
@@ -191,21 +183,26 @@ def _clahe(image, ntiles_x, ntiles_y, clip_limit, nbins=128):
xstep = x_size
xL = x - 1
xR = xL + 1
mapLU = map_array[yU, xL]
mapRU = map_array[yU, xR]
mapLB = map_array[yB, xL]
mapRB = map_array[yB, xR]
xslice = np.arange(xstart, xstart + xstep)
yslice = np.arange(ystart, ystart + ystep)
interpolate(image, xslice, yslice,
mapLU, mapRU, mapLB, mapRB, aLUT)
xstart += xstep # set pointer on next matrix */
ystart += ystep
return image
def clip_histogram(hist, clip_limit):
'''Perform clipping of the histogram and redistribution of bins.
"""Perform clipping of the histogram and redistribution of bins.
The histogram is clipped and the number of excess pixels is counted.
Afterwards the excess pixels are equally redistributed across the
@@ -213,16 +210,16 @@ def clip_histogram(hist, clip_limit):
Parameters
----------
hist : np.ndarray
histogram array
hist : ndarray
Histogram array.
clip_limit : int
maximum allowed bin count
Maximum allowed bin count.
Returns
-------
hist : np.ndarray
clipped histogram
'''
hist : ndarray
Clipped histogram.
"""
# calculate total number of excess pixels
excess_mask = hist > clip_limit
excess = hist[excess_mask]
@@ -253,28 +250,29 @@ def clip_histogram(hist, clip_limit):
hist[under] += 1
n_excess -= hist[under].size
index += 1
return hist
def map_histogram(hist, min_val, max_val, n_pixels):
'''Calculates the equalized lookup table (mapping)
"""Calculate the equalized lookup table (mapping).
It does so by cumulating the input histogram.
hist : np.ndarray
clipped histogram
hist : ndarray
Clipped histogram.
min_val : int
min value for mapping
Minimum value for mapping.
max_val : int
max value for mapping
Maximum value for mapping.
n_pixels : int
number of pixels in the region
Number of pixels in the region.
Returns
-------
out : np.ndarray
mapped intensity LUT
'''
out : ndarray
Mapped intensity LUT.
"""
out = np.cumsum(hist).astype(float)
scale = ((float)(max_val - min_val)) / n_pixels
out *= scale
@@ -285,23 +283,23 @@ def map_histogram(hist, min_val, max_val, n_pixels):
def interpolate(image, xslice, yslice,
mapLU, mapRU, mapLB, mapRB, aLUT):
'''Find the new grayscale level for a region using bilinear interpolation
"""Find the new grayscale level for a region using bilinear interpolation.
Parameters
----------
image : np.ndarray
full image
image : ndarray
Full image.
xslice, yslice : array-like
indices of the region
map* : np.ndarray
mappings of greylevels from histograms
aLUT : np.ndarray
maps grayscale levels in image to histogram levels
Indices of the region.
map* : ndarray
Mappings of greylevels from histograms.
aLUT : ndarray
Maps grayscale levels in image to histogram levels.
Returns
-------
out : np.ndarray
original image with the subregion replaced
out : ndarray
Original image with the subregion replaced.
Note
----
@@ -309,7 +307,7 @@ def interpolate(image, xslice, yslice,
within a submatrix of the image.
This is done by a bilinear interpolation between four different
mappings in order to eliminate boundary artifacts.
'''
"""
norm = xslice.size * yslice.size # Normalization factor
# interpolation weight matrices
x_coef, y_coef = np.meshgrid(np.arange(xslice.size),
@@ -325,4 +323,3 @@ def interpolate(image, xslice, yslice,
/ norm)
view[:, :] = new
return image