- Options
- Login
- Desktop Style
Options
Matching Plant Images Internally
- Matching Plant Images Internally
-
RE: Matching Plant Images Internally
Koen G > 27-09-2023, 02:08 PM
There are almost twice as many small plants as there are large plants, so it should probably not be expected that there is some neat correspondence between both groups (unless we hypothesize that a hundred or so large-plant pages are now lost).
Anyway, I am currently working on a blog post about something completely different, and I've decided to focus my research time on that first. I will tune back in here when that is finished. Meanwhile, you guys feel free to change the spreadsheet as you wish. We could also do with one or two more people rating the matches: You are not allowed to view links. Register or Login to view. -
RE: Matching Plant Images Internally
Mark Knowles > 27-09-2023, 02:41 PM
(27-09-2023, 02:08 PM)Koen G Wrote: You are not allowed to view links. Register or Login to view.There are almost twice as many small plants as there are large plants, so it should probably not be expected that there is some neat correspondence between both groups (unless we hypothesize that a hundred or so large-plant pages are now lost).
It would be my expectation that there would be twice as many small plants as large plants as there appears to be one match for the "root" and one match for the "leaf/flower". -
RE: Matching Plant Images Internally
Mark Knowles > 27-09-2023, 03:12 PM
(27-09-2023, 02:08 PM)Koen G Wrote: You are not allowed to view links. Register or Login to view.There are almost twice as many small plants as there are large plants, so it should probably not be expected that there is some neat correspondence between both groups (unless we hypothesize that a hundred or so large-plant pages are now lost).
So for example:
My matches "7", "11", "12", "13", "14", "18", "28", "34", "37" and "50" are "root" matches
My matches "24", "25", "26", "42", "44" and "48" are "leaf/flower" matches
Maybe we should have an extra column called match type which can then be: "root", "leaf/flower" or "general". (We could have "leaf" and "flower" as two separate types)
Within this framework it makes sense for there to be about double the number of small plant drawings when compared with large plants. -
RE: Matching Plant Images Internally
ReneZ > 27-09-2023, 06:13 PM
(27-09-2023, 02:08 PM)Koen G Wrote: You are not allowed to view links. Register or Login to view.unless we hypothesize that a hundred or so large-plant pages are now lost
I consider it distinctly possible that we do not have the complete set of plan representations.
Rather than lots of them being lost, perhaps many were never completed.
It is a moot point, of course, as we have no way of verifying this either way. -
RE: Matching Plant Images Internally
ReneZ > 28-09-2023, 06:34 AM
First of all thanks to Mark for sharing this work to all.
I will be happy to add to this, as soon as I have access to my normal screen again, which is next week.
Having looked at this quite a lot in the past, my experience is that this requires two screens (or one very large one). -
RE: Matching Plant Images Internally
merrimacga > 28-09-2023, 07:12 AM
(28-09-2023, 06:34 AM)ReneZ Wrote: You are not allowed to view links. Register or Login to view.First of all thanks to Mark for sharing this work to all.
I second that motion. And thanks to Koen for creating the shared spreadsheet we're using.
(28-09-2023, 06:34 AM)ReneZ Wrote: You are not allowed to view links. Register or Login to view.Having looked at this quite a lot in the past, my experience is that this requires two screens (or one very large one).
I'm finding it's possible to compare the images using the magnifier in the Folio Browser, the image comparison links for which I am working on adding to the spreadsheet. But I am using an ultrabook and I must admit that even using that magnifier - and zooming my browser to 125% - is not enough when the page, verso or recto, is one like f88v-r2. Higher than 125% zoom and the magnifier no longer accesses the whole folio. I can use the Image Extractor for greater browser zoom all the way to 500% but then the image gets a bit fuzzy and I lose the side-by-side image comparison.
I can only imagine what Mark must have had to do to create the images he added to the beginning of this thread. They are a very nice resolution.
Speaking of the Folio Browser links in the spreadsheet, though, I have hit a snag. At least one (f87v), probably two (still checking), of the folios appear to be missing from it. I informed David Jackson already and he's looking into it. -
RE: Matching Plant Images Internally
Koen G > 28-09-2023, 01:57 PM
(27-09-2023, 03:12 PM)Mark Knowles Wrote: You are not allowed to view links. Register or Login to view.Maybe we should have an extra column called match type which can then be: "root", "leaf/flower" or "general". (We could have "leaf" and "flower" as two separate types)
As I see it, the partial matches are now category 2: there are some issues, but the parts that match are intriguing enough to keep it on board.
Regarding flowers, we must also keep in mind that the small plants section seems to avoid them to a certain extent. Yes, there are flowers on some of the folios, but they really take a back seat and the focus is especially on roots, leaves and the stuff that connects them. I don't know what this means for what the section is all about, but it does imply again that we cannot expect a full correspondence between the flower-heavy large plants and the flower-averse small plants.
By the way, I have not looked at matches yet in any detail, but there were a few instances where I thought there might be a better match. For example here, I would prefer the pairing on the right over your proposal on the left:
And how to decide about something like this? Your proposal left, alternative right.
(In this case it seems pretty clear that none match, and it's really a matter of determining the least bad option). -
RE: Matching Plant Images Internally
Mark Knowles > 28-09-2023, 02:29 PM
(28-09-2023, 01:57 PM)Koen G Wrote: You are not allowed to view links. Register or Login to view.(In this case it seems pretty clear that none match, and it's really a matter of determining the least bad option).
Choosing which plants match and which don't is ultimately down to the individual. I would think that each individual will make whichever determination that they feel best serves their research and best fits their model of the author/artist's mental processes. And as stated before the closeness of each match lies on a spectrum from those which very closely match to those for which the match is very weak. And of course a researcher may reassess their matches at any time.
(28-09-2023, 01:57 PM)Koen G Wrote: You are not allowed to view links. Register or Login to view.By the way, I have not looked at matches yet in any detail, but there were a few instances where I thought there might be a better match.
You are right to look at every match in detail. Ideally what I or anyone else could do is write a note against each potential match detailing the ways in which it matches and the ways it doesn't. There can sometimes be very small, but subtle and significant differences between two plants. -
RE: Matching Plant Images Internally
Mark Knowles > 28-09-2023, 07:00 PM
(28-09-2023, 07:12 AM)merrimacga Wrote: You are not allowed to view links. Register or Login to view.
(28-09-2023, 06:34 AM)ReneZ Wrote: You are not allowed to view links. Register or Login to view.First of all thanks to Mark for sharing this work to all.
I second that motion. And thanks to Koen for creating the shared spreadsheet we're using.
The effort gone to completing the entries for the folio columns of the spreadsheet has been and is very helpful to me. I will try to assist in completing this task, but help from others is certainly very valuable. (Having the correct folio numbers for each of the matches is really useful going forward.)