(28-09-2024, 10:24 PM)Koen G Wrote: You are not allowed to view links. Register or Login to view.It's weird how it gives the nymphs a frown 
They do have a frown, and a deep one. Even if you look at visible light TIFFs from 2014, the top nymph has a frown in text ink somewhat corrected by the red paint. The frown on the right one is also there, but there is a thin stroke below it that corrects it to a kind of surprised expression. Now that I look at it carefully, there are a lot of very very unhappy nymphs there.
[
attachment=9293]
But.. look at this one. The frown almost touches the nose, where in the MS there is a bit of a whiter area. Is it picking up on erasures?
[
attachment=9294]
This is not an erasure, as far as I can see from the raw TIFFs, it's just that the model seems picking on contours around ink. So, two black lines are not the frown and the nose per se, but the area immediately below the nose and upwards from the frown. If you put the nose where the white nose shape is and the frown immediately below the black frown line, they are roughly at the same positions as with the visible light scan.
Good question is how exactly the model picks the areas surrounding ink. It won't pick visible ink, because there is no visible ink at the bottom right of You are not allowed to view links.
Register or
Login to view. where is was trained. This model processes the image in 2x2 cells, which is too small to explain these artifacts.
It's possible that at some wavelengths there is substantial glow (UV causes some luminescence) which is leaking some light into the vellum around ink spots. I've been reading and watching a bit of introductory material on MSIs, if I understood it correctly one of the reasons some hidden ink can be visible in UV because UV can reach particles of ink buried inside medium and make them glow, which lights up some small area around them.
If this is the case, the bottom part of the glow of the frown could be hidden under red paint and didn't register, that's why there is only one black line.
Have you tried mapping your greyscale images to a rainbow palette? Not sure if this improves visibility but it is worth a try.
Thank you for the suggestion, I did consider using color. However, I first thought of creating my own BW processed images mainly because it had been very hard for me to look at colorful processed images included in the MSI release. They also leave too much space for interpretation, combining color and brightness variation means 2 dimensional output for each point, which makes it very easy to start seeing things that are not actually there. My goal was to simplify.
I will try adding color to mark different features if I manage to create a model that can work reasonably well with the visible light scans.
(28-09-2024, 10:24 PM)Koen G Wrote: You are not allowed to view links. Register or Login to view.It's weird how it gives the nymphs a frown
Instantly made me sad upon seeing it.
Speaking about the width of the strokes, here's a comparison of the squiggle with the symbol from the lower right corner of f57v, roughly to scale (I used the sizes of the respective folios to match the image from the MSI set and the visible light TIFF). I'd say the width of the strokes is approximately the same. What implement was likely used for the symbol from f57v?
[
attachment=9369]
Looks like quill pen to me. The way it's cut and the amount of pressure applied can vary line thickness.