Vectorized moving window on 2D array in numpy

by

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”.

More Related Contents:

  1. Taking subarrays from numpy array with given stride/stepsize
  2. What is vectorization? [closed]
  3. Is it possible to vectorize recursive calculation of a NumPy array where each element depends on the previous one?
  4. Bin elements per row – Vectorized 2D Bincount for NumPy
  5. Vectorized searchsorted numpy
  6. Vectorized NumPy linspace for multiple start and stop values
  7. Using Numpy Vectorize on Functions that Return Vectors
  8. Vectorizing Haversine distance calculation in Python
  9. Why np.hypot and np.subtract.outer very fast compared to vanilla broadcast ? Using Numba for speedup numpy in parallel for distance matrix calculation
  10. Vectorized look-up of values in Pandas dataframe
  11. Python vectorizing nested for loops
  12. Efficient evaluation of a function at every cell of a NumPy array
  13. python recursive vectorization with timeseries
  14. How to invert a permutation array in numpy
  15. vectorize conditional assignment in pandas dataframe
  16. Subtracting numpy arrays of different shape efficiently
  17. NumPy version of “Exponential weighted moving average”, equivalent to pandas.ewm().mean()
  18. Efficiently replace elements in array based on dictionary – NumPy / Python
  19. Concatenate range arrays given start, stop numbers in a vectorized way – NumPy
  20. numpy array TypeError: only integer scalar arrays can be converted to a scalar index
  21. Numpy Broadcast to perform euclidean distance vectorized
  22. Vectorizing `numpy.random.choice` for given 2D array of probabilities along an axis
  23. Numpy: argmax over multiple axes without loop
  24. ValueError: The truth value of an array with more than one element is ambiguous. Use a.any() or a.all()
  25. How do I get indices of N maximum values in a NumPy array?
  26. Concatenating two one-dimensional NumPy arrays
  27. Roll rows of a matrix independently
  28. how does multiplication differ for NumPy Matrix vs Array classes?
  29. numpy np.apply_along_axis function speed up?
  30. Python insert numpy array into sqlite3 database

Leave a Comment