(16-03-2026, 09:25 PM)eggyk Wrote: You are not allowed to view links. Register or Login to view.It's pareidolia, trust me. It happens a lot with this type of tool. When something is really there, it's really quite obvious. You will see such things all over the page, and its often deformations in the parchment, general unevenness, or simply a part where the contrast blurs into apparent patterns which aren't there.
This is why I only tried the contrast enhancement once and You are not allowed to view links.
Register or
Login to view.. There is really no way of telling if the lines under the paint that it shows is actual ink or some artificial threshold computed from paint gradients. If some ink blobs turn into a semblance of a signature, some paint irregularities can easily turn into a resemblance of deliberate ink lines.
Uniform models that process all the pixels in the same way and do not use spacial information are much more reliable for this, but the results don't look as spectacular.
(Yesterday, 01:25 AM)oshfdk Wrote: You are not allowed to view links. Register or Login to view. (16-03-2026, 09:25 PM)eggyk Wrote: You are not allowed to view links. Register or Login to view.It's pareidolia, trust me. It happens a lot with this type of tool. When something is really there, it's really quite obvious. You will see such things all over the page, and its often deformations in the parchment, general unevenness, or simply a part where the contrast blurs into apparent patterns which aren't there.
This is why I only tried the contrast enhancement once and You are not allowed to view links. Register or Login to view.. There is really no way of telling if the lines under the paint that it shows is actual ink or some artificial threshold computed from paint gradients. If some ink blobs turn into a semblance of a signature, some paint irregularities can easily turn into a resemblance of deliberate ink lines.
Uniform models that process all the pixels in the same way and do not use spacial information are much more reliable for this, but the results don't look as spectacular.
Well, the ink blobs don't turn into any sort of signature imo. I find it very difficult to believe that these results are simply the result of paint lines. In your example the issue seemed to be lines generated from the thresholds/outlines between areas of very dark ink and the blank parchment, whereas this is highlighting slightly darker areas within an area of ink.
For example, You are not allowed to view links.
Register or
Login to view. shows a middle stem on each leaf, including the leaf behind each, where the stem is visible through the leaf spikes. I think that it would be a very big coincidence if paint lines had created the same type of stem, at the same scale, on both front leaves. And these lines aren't invisible on the regular scans either; if you zoom in and look in detail at the examples I posted you can actually make out these lines.
[
attachment=14683]
Of course, using such processing to display entire pages would be entirely inappropriate. What might work best for one area of the page may not work for another, and what might work on one colour of ink may not work work well on another.
There's a world where i'm tempted to make a composite image of each page, using different processing techniques for each aspect, or even re-imposing the processed areas back on to the original image.. but this would need to have disclaimers all over it and would only be useful for illustrative purposes. Something like this below, but done far better and with more care:
[
attachment=14684]
oeesordy wrote:
"Maybe I have Pareidolia, but could it be the signature..."
You are not seeing Pareidolia, but real and meaningful explanations, which have been deleted by someone carefully. The question is who had deleted and I have asked it without anyone being able to answer it.
The most precious is the Alphabet key on f1r- on the right rand of the foil. If you can transform it into readable - the manuscript will be deciphered.
The same writings can be found on multiple places.
The marginalia of You are not allowed to view links.
Register or
Login to view. (on the top) presents another opportunity for verification. On this place half of the text seems legible, why the rest half had been effaced
Try to apply the same transformational process on You are not allowed to view links.
Register or
Login to view. on the lowest right leaf- there too a portion is visible , another one -not.
And there are a lot of other places.
(Yesterday, 04:47 AM)BessAgritianin Wrote: You are not allowed to view links. Register or Login to view.You are not seeing Pareidolia, but real and meaningful explanations, which have been deleted by someone carefully. The question is who had deleted and I have asked it without anyone being able to answer it.
No. The example in this thread was pareidolia and to suggest otherwise in this fashion is totally irresponsible. I literally showed that it was pareidolia from the outline of ink blobs when the image was zoomed out. When zoomed in, the supposed "signature" dissapears.
At any rate, potential secret hidden signatures across the manuscript and their decoding clearly isn't the point of this thread. We're talking about leaves and image processing.
(Yesterday, 05:00 AM)eggyk Wrote: You are not allowed to view links. Register or Login to view. (Yesterday, 04:47 AM)BessAgritianin Wrote: You are not allowed to view links. Register or Login to view.You are not seeing Pareidolia, but real and meaningful explanations, which have been deleted by someone carefully. The question is who had deleted and I have asked it without anyone being able to answer it.
No. The example in this thread was pareidolia and to suggest otherwise in this fashion is totally irresponsible. I literally showed that it was pareidolia from the outline of ink blobs when the image was zoomed out. When zoomed in, the supposed "signature" dissapears.
At any rate, potential secret hidden signatures across the manuscript and their decoding clearly isn't the point of this thread. We're talking about leaves and image processing.
I believe the case I presented was indeed pareidolia BessAgritiannin, its just an artefact of the technology from detection of the ink splatter. I'm not 100% sure, but I trust eggyk analysis. What really made me confused is that Wikipedia said Jan Eyck of Liege had some secret projects. His signature is ornate, but I don't know what his penmanship is like. Pareidolia affects all we want to believe with our eyes and logic, but that can be skewed by the interpreter. Perception is a funny business but the rigors of the scientific method is our best bet so far. I was letting my opinion get in the way and I'm not a pro at this.