Undecimated discrete wavelet transform using Matlab and Python -

loosely based on above figure:

i need make array this:

array = [(image),(1,ll),(1,lh),(1,hl),(2,ll),(2,lh),(2,hl)] 

here, image numpy 2d array, , rest other array items components obtained undecimated discrete wavelet transform, 2d arrays.

for eg,

(1,ll) level 1 component,channel low-high (2,hl) level 2 component, channel high-low 

with python 2.7, simple.

udwt = np.asarray(pywt.swt2(image, 'haar',2)) array = [image,          udwt[0][0] ,    #1,ll          udwt[0][1][0],  #1,lh          udwt[0][1][1],  #1,hl          udwt[1][0] ,    #2,ll          udwt[1][1][0],  #2,lh          udwt[1][1][1]]  #2,hl 

however, unable in matlab.

udwt = ndwt2(image,2,'haar'); 

components found in udwt.dec (dec field of udwt structure), 1 7.

creating parallels python code:

python                               matlab udwt[0][0]                         question udwt[0][1][0]                      udwt.dec{5} udwt[0][1][1]                      udwt.dec{6} udwt[1][0]                         udwt.dec{1} udwt[1][1][0]                      udwt.dec{2} udwt[1][1][1]                      udwt.dec{3}  (python's function gives components level 1 level n.  matlab's function gives components level n level 1.(reverse order)) 

the problem is, not getting low-low component (approximation component)of 1st level in matlab (1,ll).

on using ndwt2 2 level decomposition, gives me 1 approximation , 3 details each level (lh,hl,hh).

whereas in python's pywt.swt2(), approximation each level, , 3 details(lh,hl,hh) each level.

so find 1,ll in matlab?

i consider udwt.dec{1} 2,ll (low-low of 2nd level instead of 1st level based on answer here.

dec{ 1 }        approximation level n 

shall using other function instead of ndwt2?