Is it a Python matlab.engine's bug for slicing 2D data?

The code is as following, when we slicing 2D data img, we've gotten unexpected result. The code is straightforward.

import matlab.engine

import matlab

img = matlab.uint8([[14,18,18,22,24,26,27,31,35,40,42,46,50,51,50,51,43,43,40,35,30,26,24,22,18],[18,19,18,24,26,30,38,43,50,58,62,70,72,78,74,72,66,58,50,43,40,34,26,26,22],[18,22,24,27,34,42,48,60,74,86,104,120,127,130,126,120,104,86,74,62,50,42,38,30,26],[22,26,26,35,46,56,67,91,118,146,174,188,200,209,199,192,167,147,112,88,70,56,43,38,27],[24,30,38,46,59,82,111,151,191,222,245,255,255,255,255,255,244,222,186,142,106,76,58,46,38],[26,34,46,58,82,116,168,221,247,255,255,255,255,255,255,255,255,255,243,202,163,112,78,56,42],[34,42,56,76,120,173,227,255,255,255,255,255,255,255,255,255,255,255,255,255,219,162,111,74,54],[40,50,67,106,170,227,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,221,155,98,66],[46,62,90,142,211,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,202,134,82],[50,70,108,184,252,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,235,165,102],[56,78,132,214,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,198,122],[59,92,155,233,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,219,142],[66,99,171,248,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,238,158],[67,106,182,254,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,243,164],[66,104,180,254,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,241,170],[66,104,177,248,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,243,159],[62,94,162,241,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,227,154],[58,86,140,220,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,208,130],[51,74,122,190,254,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,248,179,110],[46,67,98,156,222,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,212,142,88],[42,54,78,126,180,240,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,233,171,112,72],[35,43,63,90,130,186,237,255,255,255,255,255,255,255,255,255,255,255,255,255,233,174,124,78,59],[31,38,50,66,95,142,188,240,255,255,255,255,255,255,255,255,255,255,255,225,174,130,88,62,46],[26,31,42,51,70,92,127,172,210,242,255,255,255,255,255,255,248,234,200,163,120,88,63,51,40],[26,27,31,42,54,64,83,110,146,166,200,209,220,225,218,207,192,163,132,102,78,62,54,42,31]]);

print(img.size) # (25, 25), it is expected.

sub_img = img[0:14][0:14];

print(sub_img.size) # (15, 15) is expected. However, we've gotten (14, 25)

Answer 1

[Lokesh]

Hi,

I understand that the 'sub_img' obtained after slicing 2D data 'img' does not match your expected result.

When attempting to slice a 2D array in Python, and specifically a 2D MATLAB array through Python in this case, the syntax 'img[0:14][0:14]' performs the operation in two steps. Initially, 'img[0:14]' slices the array to include rows 0 through 13, as Python slicing is zero-based and the end index is exclusive. Subsequently, applying [0:14] to the result of 'img[0:14]' slices this intermediate array by rows once again, not columns, resulting in an array of size (14,25).

To correctly slice a 2D array to get a sub-array, you should use a single slicing operation that includes both rows and columns. Here is a sample code:

import matlab.engine
import matlab

# Start MATLAB engine
eng = matlab.engine.start_matlab()

# Create a sample MATLAB array
img = matlab.uint8([[14,18,18,22,24,26,27,...]]) 

# Assign the MATLAB array to the 'img' variable in the MATLAB workspace
eng.workspace['img'] = img 

# Use MATLAB's indexing (which starts at 1) and includes the end index to slice the array
sub_img = eng.eval('img(1:15, 1:15)')  # This should give you the 15x15 sub-array
print(sub_img.size)  # Should print the size of sub_img, expecting [15, 15]

Hope this helps!