- Options
- Login
- Desktop Style
Options
Matching Plant Images Internally
- Matching Plant Images Internally
-
RE: Matching Plant Images Internally
Mark Knowles > 26-09-2023, 05:26 PM
My match number "75" is an interesting example of where I have given the drawing a high match score whereas the scores of the preceding 3 scorers are low. In this case the drawings don't look identical, so one can see why they might not have been given a high score. However they look so much more alike each other than either of them do look like anything else that relative to any other plant illustrations they appear to me to be a very good match.
This is a good illustration of my approach to matching. I have tended to consider how close a match two plants are relative to other plants not just in absolute terms. -
RE: Matching Plant Images Internally
Mark Knowles > 26-09-2023, 06:06 PM
My match number "80" isn't very clear as a large part of the herbal drawing is covered, so it is probably best to refer to the original illustration. -
RE: Matching Plant Images Internally
Mark Knowles > 26-09-2023, 06:13 PM
My match number "83" is a clear example of a situation where there appears to be a leaf-root mismatch. The leaf seems to be very much the same whereas the roots are clearly completely different. This is a situation where it is very much for the researcher to determine what they think is going on here.
My interpretation is that the author/artist got confused or was careless and that explains the mismatch. Other researchers might interpret this as a deliberate mixing up of roots and leaves. -
RE: Matching Plant Images Internally
Mark Knowles > 26-09-2023, 07:37 PM
With my match "95" the herbal drawing has a root that resembles a bird. I matched it as I have as the root of the matched plant is drawn in a way that to me it also resembles a bird, although others might disagree. (It is important to note that if they are both birds they were drawn in two quite different ways.) -
RE: Matching Plant Images Internally
Mark Knowles > 26-09-2023, 07:59 PM
I have given my scores. I may change scores subject to review. I will have more to say with respect to my scoring and matching process. -
RE: Matching Plant Images Internally
Koen G > 26-09-2023, 10:23 PM
I guess it comes down to two different selection philosophies. You try to find everything which could potentially correspond in some way. I'm not saying this is bad, it is probably why you found some matches that hadn't been pointed out before.
On the other hand, I feel like it is best to start out with the more certain stuff, and then - if anything at all is found - see if this expands to other potential matches. I feel like whatever we are going to do with this is going to be complex, and we will be unable to work well with a large number of potentially bad matches. It's not like I would never consider these, but I think it's safer to start out with (the largest possible number of) clear examples. -
RE: Matching Plant Images Internally
merrimacga > 27-09-2023, 12:48 AM
(25-09-2023, 05:50 PM)Mark Knowles Wrote: You are not allowed to view links. Register or Login to view.It would be good to add two columns to the spreadsheet, one for each side of the match such as "f90r" and "f90v[1, 1]" as that information is very useful for linking rows up and then if we can all fill those in that would help.(This information is useful also for performing an "Inner Join" on the spreadsheet as we may sometimes have two [or more?] separate matches between the Herbal plant and the small botanical plants and we might want to link them all together such as a match for the root and a match for the leaf.)
(26-09-2023, 06:06 PM)Mark Knowles Wrote: You are not allowed to view links. Register or Login to view.My match number "80" isn't very clear as a large part of the herbal drawing is covered, so it is probably best to refer to the original illustration.
I have added a new column C (VN Folio Browser) to You are not allowed to view links. Register or Login to view. so links can be added to the specific images comparison in this site's Folio Browser.
To access, hover over the link in the spreadsheet then left click on the Folio Browser link in the pop up message. The specific Folio Browser comparison will then open in a new tab or new window, depending on your browser settings.
To add a link in the spreadsheet, first access the Folio Browser and then select the left and right folios to be compared and then select and copy the page URL. Then, in the spreadsheet, type the text (e.g. f48v vs f89v2 for Mark's number 9) in the column C cell, then click Enter to exit the cell. Then re-select the cell and click CTRL+K to open the pop up for adding the link. Paste the link you copied in the field in the pop up and then click the Apply link. The pop will then close. Click another cell to exit the one you edited and then the link can be accessed.
I have started adding some of the links but it will take awhile to complete. I suggest we keep columns A and B for the individual folio numbers, especially since column B also includes references to the specific small plant images in the folio. With the addition of column C, I thought it might be a good idea to freeze panes so we can scroll to the right without losing visibility on columns A-D. -
RE: Matching Plant Images Internally
Koen G > 27-09-2023, 08:34 AM
This is incredible, you are problemsolving and resurfacing forgotten forum features at the same time. -
RE: Matching Plant Images Internally
merrimacga > 27-09-2023, 09:20 AM
(27-09-2023, 08:34 AM)Koen G Wrote: You are not allowed to view links. Register or Login to view.This is incredible, you are problemsolving and resurfacing forgotten forum features at the same time.
LOL. It's how I roll!
Some of this is easy compared to the stuff I do for work (not the crypto stuff though...I have no head for EVA and ciphers and all that). This week alone I had to create a 5-page separation agreement (legal document) for an exiting employee (and a few other docs as well) plus work on updating several open enrollment spreadsheets, each with complicated formulas in a lot of the cells, adding two new medical plans and broken out amongst two employee groups with about a third of the employees replaced with new ones since the last renewal. Plus a few other tasks thrown in for good measure. And it's only Wednesday! My co-workers wanted to do dinner with me and I told them not until Thursday because I would be in a foul mood until then. A lot of the pain at work is still pandemic driven. I imagine that is true everywhere.
I wonder if the VM contains any spells or potions. Hmm... Ha! Just kidding!
All the stuff I do here gives me a break from all that and puts a smile on my face. And it feels good to help out.
-
RE: Matching Plant Images Internally
Mark Knowles > 27-09-2023, 01:42 PM
(26-09-2023, 10:23 PM)Koen G Wrote: You are not allowed to view links. Register or Login to view.I guess it comes down to two different selection philosophies. You try to find everything which could potentially correspond in some way. I'm not saying this is bad, it is probably why you found some matches that hadn't been pointed out before.
On the other hand, I feel like it is best to start out with the more certain stuff, and then - if anything at all is found - see if this expands to other potential matches. I feel like whatever we are going to do with this is going to be complex, and we will be unable to work well with a large number of potentially bad matches. It's not like I would never consider these, but I think it's safer to start out with (the largest possible number of) clear examples.
The way I produced my list was by comparing every herbal plant drawing with each of the small botanical drawings and listing the 4 best matches. I then compared those matches to see if there was one that stood out. If there was one match that stood out from the others and appeared to me a match then I would record that. Of course, in the case of some herbal plant drawings there was no match, but in the case of others there was one which is to be found in my list.
There is definitely scope for different approaches. I have tended to work on the basis that it is reasonable to expect most herbal plants will also appear in the botanical section as that would seem to make sense in the context of recipes. So this makes me more inclined to match two drawings. I think it makes sense to treat some matches as "fuzzy" matches, i.e. matches that may in fact not be matches after all, so they can be excluded or included if it seems appropriate.