Calculating variance image python
有没有一种简单的方法可以使用 Python/NumPy/Scipy 计算图像上的运行方差过滤器?通过运行方差图像,我的意思是计算图像中每个子窗口 I 的 sum((I – mean(I))^2)/nPixels 的结果。
由于图像非常大(12000×12000 像素),我想避免在格式之间转换数组的开销,以便能够使用不同的库然后再转换回来。
我想我可以通过使用类似
的方法找到平均值来手动执行此操作
1
2 3 4 5 |
kernel = np.ones((winSize, winSize))/winSize**2
image_mean = scipy.ndimage.convolve(image, kernel) diff = (image – image_mean)**2 # Calculate sum over winSize*winSize sub-images # Subsample result |
但是如果有像 Matlab 中的 stdfilt 函数这样的东西会更好。
任何人都可以指出具有此功能并支持 numpy 数组的库的方向,或者提示/提供一种在 NumPy/SciPy 中执行此操作的方法吗?
更简单的解决方案也更快:使用 SciPy//’s
1
2 3 4 5 6 7 8 9 |
import numpy as np from scipy import ndimage rows, cols = 500, 500 win_rows, win_cols = 5, 5 img = np.random.rand(rows, cols) |
“跨步技巧”是个漂亮的技巧,但速度慢了 4 且不那么可读。
您可以使用
1
2 3 4 5 6 7 8 9 10 11 |
import numpy as np from numpy.lib.stride_tricks import as_strided rows, cols = 500, 500 img = np.random.rand(rows, cols) win_img = as_strided(img, shape=(rows-win_rows+1, cols-win_cols+1, |
现在
1
2 3 4 5 6 7 8 9 10 11 12 |
>>> img[100:105, 100:105]
array([[ 0.34150754, 0.17888323, 0.67222354, 0.9020784 , 0.48826682], [ 0.68451774, 0.14887515, 0.44892615, 0.33352743, 0.22090103], [ 0.41114758, 0.82608407, 0.77190533, 0.42830363, 0.57300759], [ 0.68435626, 0.94874394, 0.55238567, 0.40367885, 0.42955156], [ 0.59359203, 0.62237553, 0.58428725, 0.58608119, 0.29157555]]) >>> win_img[100,100] array([[ 0.34150754, 0.17888323, 0.67222354, 0.9020784 , 0.48826682], [ 0.68451774, 0.14887515, 0.44892615, 0.33352743, 0.22090103], [ 0.41114758, 0.82608407, 0.77190533, 0.42830363, 0.57300759], [ 0.68435626, 0.94874394, 0.55238567, 0.40367885, 0.42955156], [ 0.59359203, 0.62237553, 0.58428725, 0.58608119, 0.29157555]]) |
但是,您必须小心,不要将图像的窗口视图转换为它的窗口副本:在我的示例中,这将需要 25 倍的存储空间。我相信 numpy 1.7 可以让您选择多个轴,因此您可以简单地执行以下操作:
1
|
>>> np.var(win_img, axis=(–1, –2))
|
我被 numpy 1.6.2 卡住了,所以我无法测试它。如果我没记错我的数学,另一个选项可能会因窗口不太大而失败:
1
2 3 |
>>> win_mean = np.sum(np.sum(win_img, axis=–1), axis=–1)/win_rows/win_cols
>>> win_sqr_mean = np.sum(np.sum(win_img**2, axis=–1), axis=–1)/win_rows/win_cols >>> win_var = win_sqr_mean – win_mean**2 |
现在
1
2 |
>>> win_var.shape
(496, 496) |
和
经过一些优化,我们想出了一个通用 3D 图像的函数:
1
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 |
def variance_filter( img, VAR_FILTER_SIZE ): from numpy.lib.stride_tricks import as_strided WIN_SIZE=(2*VAR_FILTER_SIZE)+1 # Check that there is a 3D image input. # Set up a windowed view on the data… this will have a border removed compared to the img_in # Calculate variance, vectorially # As per http://en.wikipedia.org/wiki/Variance, we are removing the mean from every window, # Prepare an output image of the right size, in order to replace the border removed with the windowed view call # output |
您可以使用
1
2 3 4 5 |
import numpy as np
import scipy.ndimage as ndimage subs = 10 # this is the size of the (square) sub-windows img = np.random.rand(500, 500) img_std = ndimage.filters.generic_filter(img, np.std, size=subs) |
您可以使用
原创文章,作者:ItWorker,如若转载,请注明出处:https://blog.ytso.com/tech/python/268043.html