Closed Claudius-Appel closed 2 months ago
This would allow saving the masks for each spectrum and each image.
However, this is quite expensive computationally/time-wise:
$N\text{total-masks}=k{images}\times n{spectra}\times b\text{boolean normal or extreme}$
For 32 images and normal GFA-procedure with 4 spectra, that results in
$N = 32 4 2 = 256$
images at 6000x4000 pixel resolution. Each image must be generated manually, sequentially, and thus this will take ages.
I suppose this could be parallelized across images, the same way the main analysis pipeline is - but idk.
Spectra can also not be masked collectively in a single image as they might overlap (e.g. bex_drought_HSV and bex_complete_HSV)
bex_drought_HSV
bex_complete_HSV
This would allow saving the masks for each spectrum and each image.
However, this is quite expensive computationally/time-wise:
$N\text{total-masks}=k{images}\times n{spectra}\times b\text{boolean normal or extreme}$
For 32 images and normal GFA-procedure with 4 spectra, that results in
$N = 32 4 2 = 256$
images at 6000x4000 pixel resolution. Each image must be generated manually, sequentially, and thus this will take ages.
I suppose this could be parallelized across images, the same way the main analysis pipeline is - but idk.