Welcome to OGeek Q&A Community for programmer and developer-Open, Learning and Share
Welcome To Ask or Share your Answers For Others

Categories

0 votes
299 views
in Technique[技术] by (71.8m points)

python - Vectorized moving window on 2D array in numpy

I am apply an operation on a moving window of constant size across a 2D array. Is there an efficient vectorize-like operation I can implement to do this without looping in Python? My current structure looks something like this

 for i in range(1,xmax-1):
     for j in range(1,ymax-1):
        out[i][j] = f(in[i][j],in[i+1][j],in[i-1][j],in[i][j+1],in[i][j-1],...)

The comments that eat left in this question allude to the possibility of vectorizing this operation this, but without further details vectorized indexing/slicing in numpy/scipy?

See Question&Answers more detail:os

与恶龙缠斗过久,自身亦成为恶龙;凝视深渊过久,深渊将回以凝视…
Welcome To Ask or Share your Answers For Others

1 Reply

0 votes
by (71.8m points)

You can use the rolling window technique as explained here, here and here, but for 2D array.

The source code for 2D rolling window in NumPy:

# Rolling window for 2D arrays in NumPy
import numpy as np

def rolling_window(a, shape):  # rolling window for 2D array
    s = (a.shape[0] - shape[0] + 1,) + (a.shape[1] - shape[1] + 1,) + shape
    strides = a.strides + a.strides
    return np.lib.stride_tricks.as_strided(a, shape=s, strides=strides)

a = np.array([[0,  1,  2,  3,  4,  5],
              [6,  7,  8,  9, 10,  11],
              [12, 13, 14, 15, 7,   8],
              [18, 19, 20, 21, 13, 14],
              [24, 25, 26, 27, 19, 20],
              [30, 31, 32, 33, 34, 35]], dtype=np.int)
b = np.arange(36, dtype=np.float).reshape(6,6)
present = np.array([[7,8],[13,14],[19,20]], dtype=np.int)
absent  = np.array([[7,8],[42,14],[19,20]], dtype=np.int)

found = np.all(np.all(rolling_window(a, present.shape) == present, axis=2), axis=2)
print(np.transpose(found.nonzero()))
found = np.all(np.all(rolling_window(b, present.shape) == present, axis=2), axis=2)
print(np.transpose(found.nonzero()))
found = np.all(np.all(rolling_window(a, absent.shape) == absent, axis=2), axis=2)
print(np.transpose(found.nonzero()))

Array present is occurred in array a two times on [1,1] and [2,4].

More examples in my CoLab notebook "Rolling window on NumPy arrays without for loops".


与恶龙缠斗过久,自身亦成为恶龙;凝视深渊过久,深渊将回以凝视…
OGeek|极客中国-欢迎来到极客的世界,一个免费开放的程序员编程交流平台!开放,进步,分享!让技术改变生活,让极客改变未来! Welcome to OGeek Q&A Community for programmer and developer-Open, Learning and Share
Click Here to Ask a Question

...