You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Documentation on Interpolations for imshow and Image antialiasing #20494
I found interpolations for imshow and image antialiasing to be unclear because for natural images it is fair to use hanning interpolations due to the frequency spectrum for natural images but for something like a matrix of random Bernoulli this can lead to aliasing effects and other interpolations schemes such as nearest would be more appropriate. In general, this also gives a good example for why we should use matshow instead for imshow to visualise matrices.
As pointed out this would be closed by #18749 and #18782
(Code can be included in GitHub pages directly. Please do not post screenshots.) I'll close here as we are already aware of this concern and tracking it.
Activity
jklymak commentedon Jun 23, 2021
Please include code in issues, its hard to tell what
Xis.If the anti-aliasing is kicking in, you do not have enough resolution to see any details.
There is a documentation draft at #18749 explaining the difference between data interpolation and optical anti-aliasing.
There is a PR at #18782 to add optical anti-aliasing. There is a pretty firm plan to make that available for 3.5.
JamesMLdoesML commentedon Jun 23, 2021
Hi @jklymak, X is a random matrix with Bernoulli distributions where p=0.5
jklymak commentedon Jun 23, 2021
As pointed out this would be closed by #18749 and #18782
(Code can be included in GitHub pages directly. Please do not post screenshots.) I'll close here as we are already aware of this concern and tracking it.