python - Determine adjacent regions in numpy array -

i looking following. have numpy array labeled regions. numpy array represents segmented image. region number of adjacent cells same value. each region has own unique value. simplified version 3 regions this:

x = np.array([[1, 1, 1], [1, 1, 2], [2, 2, 2], [3, 3, 3]], np.int32) 

output:

array([[1, 1, 1],        [1, 1, 2],        [2, 2, 2],        [3, 3, 3]]) 

in above example have 3 separate regions, each labeled unique value (1,2,3 in case).

what want value of adjacent (neighbor) regions each individual region. in case:

• region 1 adjacent region 2
• region 2 adjacent region 1 , 3
• region 3 adjacent region 2

what elegant , fastest way of achieving this?

many thanks!

i understand task return distinct entries of array adjacent given number (such 2). 1 way achieve numpy methods use roll shift given region 1 unit up, down, left, , right. logical or of shifted regions taken, , distinct elements match condition returned. remains remove region itself, since it's not considered own neighbor.

since roll re-introduces values move beyond array's bounds @ opposite ends (which not desired here), additional step replace row or column false.

import numpy np  x = np.array([[1, 1, 1], [1, 1, 2], [2, 2, 2], [3, 3, 3]], np.int32) region = 2   # number of region neighbors want  y = x == region  # convert boolean  rolled = np.roll(y, 1, axis=0)          # shift down rolled[0, :] = false              z = np.logical_or(y, rolled)  rolled = np.roll(y, -1, axis=0)         # shift  rolled[-1, :] = false z = np.logical_or(z, rolled)  rolled = np.roll(y, 1, axis=1)          # shift right rolled[:, 0] = false z = np.logical_or(z, rolled)  rolled = np.roll(y, -1, axis=1)         # shift left rolled[:, -1] = false z = np.logical_or(z, rolled)  neighbors = set(np.unique(np.extract(z, x))) - set([region]) print(neighbors)