search

colinr23 / spatialcompare

methods for comparing spatial patterns
5 stars 1 forks source link

weighted values appear off for scenario analysis #5

Open colinr23 opened 6 years ago

colinr23 commented 6 years ago

Replicating table 1 in Maruca and Jacquez does not get correct values in all cases. Weighted results appear to be off, perhaps due to area functions noted in other issues

jasperhyp commented 3 years ago

Hi, I am not sure if there's some issues regarding the implmentation. Please refer to my stackoverflow question here: https://stackoverflow.com/questions/66346949/why-ssim-implementations-by-different-r-packages-yield-drastically-different-res. Thank you!

colinr23 commented 3 years ago

Hi - havent looked at this codebase for a while, but we did validate against the original wang numbers on th einstein benchmark dataset: https://github.com/colinr23/spatialcompare/blob/master/vignettes/spatialcompare-vignette.pdf

https://ece.uwaterloo.ca/~z70wang/research/ssim/

Best, Colin

On Wed, Feb 24, 2021 at 1:07 AM jasperhyp notifications@github.com wrote:

Hi, I am not sure if there's some issues regarding the implmentation. Please refer to my stackoverflow question here: https://stackoverflow.com/questions/66346949/why-ssim-implementations-by-different-r-packages-yield-drastically-different-res. Thank you!

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/colinr23/spatialcompare/issues/5#issuecomment-784924680, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAN4Y2YS7ILUDQVWUQZ376TTAS6W7ANCNFSM4DWK3IKA .