Hey there
I am wondering if I am doing something wrong when running the QPCA framework, specifically the RNL ranked filter.
I have a RAW image (6984 x 4660 pixels). I generate a multispec image and then create two ROIs – y animal and the background. The background isn’t a huge area around the bird – but enough to capture the background. I then create a cone catch model, and because it is for human vision also create a luminance channel (lw and mw).
I am then running the QPCA framework (just acuity modelling and the RNL Ranked filter for now) on the animal ROI and the background ROI independently. When I run the animal ROI using the Gaussian filter it runs totally fine but when I run the background, it is taking me over an hour. I know that the Gaussian filter takes longer than AcuityView but this really seems to be a long time. I’ve noticed that the image doesn’t seem to scale down and is 1993 x 1204 pixels – I am assuming this is what is causing the long processing times.
I have followed the online tutorials to a tee and have noticed on the online videos that the images seem to rescale smaller quite significantly. Perhaps this is quite normal but I can’t help but feel that there is something that I am missing something (and I definitely rescale to 5 px per MRA).
Any help you can give would be really appreciated.
Cheers, Lou
Hi,
If the image does not appear to be noticeably scaling down in terms of spatial content, then the spatial resolution of your visual system at that particular viewing distance can resolve that much. Given the change of resolution, it would seem to be downscaling significantly, nevertheless, as would be expected. You are right that the size of the image or ROI is what determines the duration of any processing steps.
Cheers,
Cedric