Migrate metadata from ACDsee images for use in IMatch

[Bones]

I am trialing migrating from ACDSee to Imatch to use IMAnywhere and would like to use the data embeded into the images (rather than via CSV etc as i suspect the applications will both be running for a while while i learn all (or enough) aspects of IMatch...)

I see in Metadata Browse the XMP-acdsee data - particularly Category, Keywords & Collections (i have already used Exiftool to migrate the face regions from acdsee to MWG successfully). These are all structured information (xml or seperated style - see image 2 attached)

I have read in the help and forums that @Keywords can create Categories from Keywords (this has only seemed to partially worked sofar  - only picking up XMP Adobe Lightroom Hierarchical Keywords - so i am not fully understanding functionality yet)

i see other concepts that may apply - @Builder & Templates, Calling Exiftool on scan/rescan - possible?

I would appreciate anyone elses experience on if what i am atempting is resonably possible or the best way to approach it or how have others achieved simmilar things - maybe from other applications.

I will continue reading and trying as i can reaily see that the flexibility of IMatch is incredibale and will take me some time to get a handle on

The two screenshots below show - for an Image that had ACDSee Categories, Keywords and Collections saved into IPTC Keywords

• 1. what  keywords wer loaded into IMAtch - ie C&H Engage, Coutney, Hewy, Jannine, Steve,  D-All Good Pics, E-2021 03 Courtney & Hugh Engage
•    
   with only  D-All Good Pics, E-2021 03 Courtney & Hugh Engage anding up in @Keywords
• 2. the xmp-acdse metadata recognised that i would like to load into IMatch...

Any input appreciated

[Mario]

I have read in the help and forums that @Keywords can create Categories from Keywords (this has only seemed to partially worked sofar  - only picking up XMP Adobe Lightroom Hierarchical Keywords - so i am not fully understanding functionality yet)

This is how this works. IMatch internally always works with hierarchical keywords because they convey the most information. It automatically maps hierarchical keywords to flat XMP keywords and optionally existing legacy IPTC keywords.

I recommend reading Metadata for Beginners, because this topic explains this this and a lot more.

IMatch does not care for proprietary ACDSee metadata this application has stored in your files.
The data will be imported and is available for sorting, searching, editing, data-driven categories like all other XMP data. But IMatch will not work with the proprietary ACDSee keywords or ACDSee categories or collections. I have no idea what this metadata does and I have never used ACDSee (I have IMatch, so...).

By the looks of it, the tags contain structured data and ExifTool has converted that structure into XMP - this is format IMatch uses when communicating with ExifTool.
You cannot directly deal with XMP structures in IMatch. Usually these are hidden in all standard layouts, except in the "Browser" layout you have used, which shows all tags. Do not modify these tags!
ExifTool will de-structure these tags into individual tags, which you can use in IMatch for variables and all other purposes.

I see a tag named caption and rating in the ACDSee namespace. Which also seems to appear in the legacy IPTC record you show in your screen shot.

Unfortunately your screen shot does not show the important metadata in the XMP namespaces, IPTCCore and IPTCExt.
This would show us how good ACDSee maintained the official standard XMP metadata in addition to their own custom metadata.

If ACDSee maintains the data in official XMP tags, you can use it immediately in IMatch. And the proprietary ACDSee namespace becomes more or less ballast, unless you want to keep some of the information it contains.

Use the "Default" Metadata Panel layout, which covers all frequently used XMP fields.
Use the "IPTCore/IPTCExt" layout to see all standard IPTC tags embedded in your images or produced by IMatch during import from native EXIF/GPS data.

Note: most of the tags in the IPTC Application Record in your screen shot are read-only and marked as linked (see the icon) because they are linked to an XMP tag and are automatically updated by IMatch during write-back when the target file has legacy IPTC data. You don't work with these tags directly, only their XMP counterparts.

See Metadata for Beginners for more information about.

Does the Keywords Panel show proper hierarchical keywords?
During import, IMatch looks for existing standard hierarchical keywords and flat keywords in XMP and legacy IPTC.
It merges these keywords produces hierarchical keywords as necessary, based on the thesaurus you have created. See Keyword Mapping during File Import for details. The defaults used by IMatch usually just work.

For example: There is a flat keyword named "London" and your thesaurus has a hierarchical keyword "LOCATION|City|London". IMatch then maps the flat London keyword into the hierarchical keyword LOCATION|City|London during import automatically.

If ACDSee did not fill the standard and commonly used XMP tags, XMP::dc:Subject and XMP::lr:hierarchicalSubject etc., you can copy the data from the proprietary ACDSee namespace into the standard tags.

IMatch offers tools like the Metadata Mechanic which makes it easy to copy metadata from one tag to another (in your case from a proprietary ACDSee tag into the proper standard XMP tag). Metadata Templates can also be used, if the copying has to be done on multiple occasions or in steps.

Creating data-driven Categories from on proprietary ACDSee metadata may be helpful during the initial organization.

Please send me a couple of the images with the proprietary ACDSee XMP data (see support email address). I can have a look when there is time and maybe add some related tips to the Metadata for Beginners help topic.

[Mario]

Thanks for sending two sample images with metadata produced by ACDSee.
I must say I'm not overly impressed.

ACDSee wrote a small number of the standard XMP tags:

Code Select Expand

[XMP-xmp]      Rating
[XMP-xmp]      Label
[XMP-photoshop] Instructions
[XMP-photoshop] City
[XMP-photoshop] State
[XMP-photoshop] Country
[XMP-photoshop] Transmission Reference
[XMP-photoshop] Credit
[XMP-photoshop] Source
[XMP-photoshop] Headline
[XMP-iptcCore]  Location
[XMP-iptcCore]  Country Code
[XMP-iptcCore]  Creator Work URL
[XMP-xmpRights] Usage Terms
[XMP-dc]        Description
[XMP-dc]        Subject
and has 'buried' all other information in the proprietary ACDSee XMP namespace extension.

They have written flat keywords (without hierarchy) into the XMP-dc:Subject tag. IMatch picks them up from there, but without a thesaurus to map them into a hierarchy, they end up as flat keywords in keywords.

I see a proprietary tag named XMP-acdsee:Keywords  which seems to contain hierarchical keywords (?) of sorts, since there are values like "Albums|D-All Good Pics".

The existing keywords in the file are (names changed to initials for privacy):

XMP-dc:Subject (official XMP keyword tag):

C&H Engage;
C;
Hy;
J;
S;
D-All Good Pics;
E-2021 03 C & H Engage

ACDSee (hierarchical) Keywords

Albums|E-2021 03 C & H Engage
Albums|D-All Good Pics

There are 3 keywords which only exists in the flat XMP keywords but not in the ACDSee keywords: Hy, J, S.
I have no idea why these are not also in the ACDSee keywords.

When IMatch imports the file, it only looks at the official XMP-dc:Subject keyword tag and imports it into hierarchical keywords. Since there is no thesaurus that has an entry like Albums|D-All Good Pics  or anything, IMatch has no way to map the flat keywords written by ACDSee into a hierarchy. So we end up with

C&H Engage;
C; D-All Good Pics;
E-2021 03 C & H Engage;
Hy;
J;
S

One way to deal with that would be to setup keywords in the IMatch Thesaurus to match the keyword layout you have created in ACDSee (?) Maybe this can be exported in one of the many formats supported by the IMatch Thesaurus. When you then re-import the metadata, IMatch would import the keywords into a proper hierarchy.

I've made a quick test with a Metadata Template, for importing the ACDSee keywords into hierarchical keywords. It's very simple:



When I run it, we end up with two proper hierarchical keywords:

Albums|E-2021 03 C & H Engage
Albums|D-All Good Pics

and the automatically created @Keywords Categories:

Albums
 |-- E-2021 03 C & H Engage
 |-- D-All Good Pics

which is nice.
Unfortunately, since 3 of the keywords ACDSee has written to the flat keywords in XMP-dc:Subject are not in the ACDSee keywords tag, they are lost that way.

When can set the Metadata Template to append the keywords from ACDSee to the flat keywords IMatch has imported from the XMP-dc:Subject tag, but that would duplicate some of the keywords and some manual cleanup is required.

The best solution seems to be to re-create the Thesaurus/Keyword Hierarchy you have in ACDSee in the IMatch Thesaurus and then to re-import the images so IMatch can map the keywords into a proper hierarchy.

Face Regions

I see tags named like [XMP-acdsee-rs] Regions Applied To Dimensions which seem to contain face regions.
Unfortunately, ACDSee again made their own thing instead of using the official face region tags in XMP.
IMatch is hence unable to automatically import the face regions and tags (see Working with XMP Face Regions for reference).

Probably this could be resolved with some ExifTool trickery or Metadata Templates or the Metadata Mechanic, but I don't have the time.

If ACDSee would have used the official standard XMP tags, and also filled the proprietary but now industry standard XMP-lr:hierarchicalSubject tag, this would have been very easy.