[Octave-bug-tracker] [bug #62452] [octave forge] (image) bwmorph spur produces wrong result

[Hartmut]

Follow-up Comment #21, bug #62452 (project octave):

I have tested the code change of applylut.m, as mentioned in comment #20. This
code change acutally means to rotate the filter kernel by 180 degrees. 

Unfortunatelly it does not seem to be that easy. After doing this code change,
the first unit test ("2-by-2 test") in applylut.m fails. The result there
should be [1 1 1], this is also what Matlab gives, but with the suggested code
change, the new Octave result becomes [1 1 0] and therefore the unit test in
applylut.m rightfully fails.

I have not checked if this code change (comment #20) improves anything in
bwmorph, as intended. (I expect it will.) But I think we canNOT do this code
change because it would generate a regression in applylut, and probably
generate much more issues than it solves.


    _______________________________________________________

Reply to this item at:

  <https://savannah.gnu.org/bugs/?62452>

_______________________________________________
Message sent via Savannah
https://savannah.gnu.org/