Categories and Keywords

[TomS]

I'm getting ready to trial iMatch but haven't actually installed it yet. I have started reading the help file first. I am a little confused by the relationship between Keywords and Categories. It seems to me that I will end up with many Categories that match my keywords. Let me try to use an example to explain what I mean.

Let's say that I have pictures of swallowtail butterflies and I want to have a category for Butterfly with a child category of Swallowtail with a child category named Eastern Tiger Swallowtail (among others on the same hierarchical level). Now I also want to write to my file the keywords "Butterfly" and "Eastern Tiger Swallowtail" but not "Swallowtail". It seems to me that I would have to create multiple hierarchies to accomplish this, one would be a hierarchical keyword Butterfly|Eastern Tiger Swallowtail and the other would be categories with the structure Butterfly|Swallowtail|Eastern Tiger Swallowtail. Is this correct? Is there a better way to do this? It seems like I'll be duplicating a lot of effort this way.

I guess what I'm getting at is that I want to have hierarchical categories in which some of the levels would have corresponding keywords and others would not. If I assign a category to a file (or vice versa) I would like the keywords to be assigned at the same time. As in my example above, I would like to assign a file to the  Eastern Swallowtail Butterfly "category" and have the keywords "Butterfly" and "Eastern Tiger Swallowtail" written to my file as flat keywords in the process. Is that possible or would I have to assign my file to the category and then add those keywords to my file in a separate step?

[Aubrey]

Hello Tom,
Welcome to the IMatch forum! 
My thoughts on the keywords...
Why would you have "Butterfly" and "Eastern Swallowtail" as separate keywords. If you place "Butterfly Eastern Swallowtail" in your keywords, then this will cover both for any keyword search

For example I have a Rice Paper butterfly in my categories. My population of categories to keywords is:
Locations|America North|USA|Texas|Houston|Museum of Natural Science|Cockrell Butterfly Center; Nature|Butterflies|Rice Paper

In general my categories is based on Where, Who, What, When
So my categories in this case are based on Where and What

Based on your categories you would probably get Butterfly|Eastern Swallowtail in your keywords if populated from your categories using IMatch menu Command|Meta-templates|Category to Keyword

Aubrey.

[TomS]

Thanks Aubrey,

I think what I was really looking for was the group level elements and excluded elements. I finally found those in the help file. I really need to do some experimentation with the software before I can ask more intelligent questions. I'm sorry for asking these questions now before investing more time experimenting. I'm just trying to prepare myself for the trial since it appears that iMatch will have a pretty steep learning curve and I don't know how much time I'll actually have over the next 30-days to experiment.

[DigPeter]

Tom - Keywords are embedded in the metadata of a file.  They can be identical to Categories, which are contained in IMatch's database.  @Keywords are a form of category which is generated from the keywords in the file's metadata.  Many people keep a list of all relevant keywords in a thesaurus for ease of reference.

I use hierarchical keywords, which I think addresses you problem  In the case you mention, I have a the following hierarchy:

Lepidoptera
      Butterflies
         species (e.g. Eastern Tiger Swallowtail)
      Moths
         species

So your butterfly's keyword would look like Lepidoptera|Butterflies|Eastern Tiger Swallowtail   This would be in my thesaurus, so in the keyword panel, I mark this in the thesaurus and it is assigned to the file.  Depending on the preferences which you have set, IMatch then takes control and writes this into the file's metadata as a hierarchical keyword and also as the individual keywords "Lepidoptera", "Butterfly" and "Eastern Tiger Swallowtail".  IMatch also writes this as a data -driven category to the database, so that it appears in the Category view and panel.

[Mario]

Categories and Keywords are in fact separate things.

+ Keywords are stored as part of the metadata inside your image files. Many applications support keywords.

+ Categories are stored in your database, offer a lot more functionality than keywords, and also work for file formats which don't support embedded keywords.

The special @Keywords category basically mirrors the keywords contained in your files. This makes it easy to quickly see files with one or more keywords, or to use keywords stored in your files like normal IMatch categories, e.g. in formulas or with the category @Builder.

@Keywords has some extra features specifically designed to make working with keywords easier, e.g. adding/removing keywords via drag & drop, visualize keyword hierarchies (e.g., for scientific taxonomies) etc.

The main feature in IMatch to work with keywords is the Keyword Panel. Here you see and edit the keywords in your files.
When you add, change or remove keywords in this panel, these changes are stored in the metadata in your files, and are automatically mirrored in the @Keywords category IMatch maintains for you.

The thesaurus is the feature in IMatch where you can optionally setup thesaurus data, e.g. a controlled vocabulary or a taxonomy for keywords. The thesaurus is displayed in the Keyword Panel and allows you to quickly pick keywords for your files, without the need to type them in. Using the thesaurus makes input quicker, avoids typos and improves quality. The thesaurus offers features like synonyms and enables you to assign entire groups of keywords with just one click. All this has been designed to reduce the time spent adding keywords while improving the quality and consistency.

Whatever you do in the Keywords Panel, the keywords will finally end up in your files, and the @Keywords category will mirror them in order to offer you additional functionality.

[TomS]

Thank you all for your responses. I do understand the difference between Categories and Keywords in IMatch5. What was confusing me was that I wanted to find ways to create a hierarchy of keywords in which some of the levels of that hierarchy were used for grouping only and would not be written to the file as flat keywords. I believe I've found how to accomplish that in the help file but have not yet tried it.

One additional question I have, which I will answer myself eventually when I finally create my test database and import a few files, is how will IMatch respond to a file that has only one flat keyword in the dc:subject tag - "Black Swallowtail Butterfly" but has a hierarchy in the lr:hierarchicalSubject tag - "Object|Animal|Insect|Butterfly|Swallowtail|Black Swallowtail Butterfly"? This is just one example and I realize it may not be what others would do but it is what is currently in my file so I need to handle it.

Basically, when I tagged this image before, I created a hierarchy but specified that I only wanted to create a keyword for the lowest level element in that hierarchy. The software I used to do this was configured to write hierarchical keywords to the lr:hierarchicalSubject tag in addition to writing flat keywords to the dc:subject tag, which isn't really any different that what IMatch does. The software I used allowed me to specify in each element of my hierarchy whether or not to create flat keywords for them. I'm curious about how this will be reflected in the IMatch database once the file is imported.

One could easily debate whether or not doing what I did makes sense and I am still questioning some of my keywording strategy but I have about 30,000 images to work with, which already contain keywords and other metadata that was added so I want to know what to expect if I start importing bunches of them into IMatch.

[Mario]

What was confusing me was that I wanted to find ways to create a hierarchy of keywords in which some of the levels of that hierarchy were used for grouping only and would not be written to the file as flat keywords. I believe I've found how to accomplish that in the help file but have not yet tried it.

You can control this via the IMatch thesaurus. If allows you to flag levels as 'group levels'. Such levels are trimmed when IMatch maps hierarchical keywords into flat keywords.

When IMatch imports flat keywords (e.g., during the initial import) and you configure IMatch to lookup keywords via the thesaurus (Edit > Preferences > Metadata) IMatch considers group levels as well to match flat keywords to hierarchical keywords.

One additional question

IMatch imports existing flag keywords into hierarchical keywords on import. You thus will get a 'marge' if you have used different keywords in IPTC, XMP and hierarchical keywords. IMatch manages one set of keywords per file (hierarchical) and this is what you see and work with. When keywords are written to the file, IMatch synchronizes hierarchical keywords into flat XMP and optionally legacy IPTC keywords.

I strongly suggest you first install IMatch, work with it for a few days, test how it behaves with your files etc. This will answer many of the questions you might have. Start with a test database which you may (or may not) throw away later. If you want to be on the super-safe side, make a copy of some folders containing images and only work with these duplicates in IMatch. This way, you can start over at any time and try anything out without risking anything.

[Panther]

I hope this isn't considered hijacking this thread, but reading these posts has raised a question related to the use of and relationship between categories and keywords that I was hoping might have a simple answer that's related to what has already been said here.

So far, I have done everything about organizing my image files in IMatch using categories only, and have done nothing with keywords.  After reading this thread it struck me that maybe it wouldn't be a bad idea to also be using keywords since apparently they are actually saved in/with the image files themselves instead of just in the IMatch database.

Is there a simple way to tell IMatch to add keywords to every image file equal to the categories I have assigned to that image file, without having to do things manually for each file, and is there any reason that wouldn't be a good idea?

[Mario]

You can do that using a metadata template.
Basically you use your categories (or a subset) as the source (via  a variable) and the hierarchical as the target.
I'm sure we had a number of threads about this here. Try to search and if you don't find it, holler for more details.

Note: Not all file formats support keywords. Updating categories is super-fast, writing out the keywords to the files can be slow, especially if you use large files. Try first with some sample files to see how it performs.

Persisting keywords in files is usually a good idea, because it allows you to use them in other applications too. That's why IMatch supports a two-way workflow for categories and keywords. Keywords are 'dumb', compared to categories. But IMatch smartens them up with the @Keywords category. This gives you the best of both worlds.

[tmcgill]

After reading this thread it struck me that maybe it wouldn't be a bad idea to also be using keywords since apparently they are actually saved in/with the image files themselves instead of just in the IMatch database.

Indeed, and I've wondered the same ever since IMatch 5 came out and enhanced keyword functionality to the point that I could use it to accomplish a significant portion of what I use categories for. However, there is one important thing to remember here: if your image categorization scheme in any way conveys personal or proprietary information, switching over to keywords would require you to always be vigilant any time you ever share an image from your library with anyone else, especially publicly. By embedding the info in the file itself, you are sharing the entire classification scheme you've applied to an image with anyone who sees that image, unless you go through the multi-step process of copying your images somewhere else, then stripping all the hierarchical keywords out, then sharing the images. I'm sure for some people this is already part of their workflow, but I haven't taken the step of implementing that kind of thing yet, since I haven't had any strong need up to this point to make my categories exist outside my IMatch database.

[jch2103]

You're certainly right about the trade-offs of using categories v keywords. However, note that iMatch does make it easy to strip metadata out when using the Image Batch Processor. A preset can define which (if any) metadata is included, as well as specific destination, etc. This may not be what you'd want to do, but it is accounted for in iMatch.

[tmcgill]

Right, hopefully I didn't imply it would be very hard to do that, just that it was an important set of steps that you might want to make sure to do if you use keywords that way. (Can't just go dragging your files straight out of your library onto Facebook or Flickr or onto a CD to give your client, or wherever you're putting them, if you have things embedded in the files like names and addresses of family members or customers, or other info you might want to keep private.)

[Mario]

IMatch gives you many options to deal with such questions and privacy issues.

To name a few:

+ You can decide which information you want to keep in keywords, and which only in categories.
+ You can utilize your keywords like categories using the special @Keywords category.
+ You can use the Batch Processor to produce "clean" outputs without only limited metadata.
+ You can just copy your files in IMatch and then run an ExifTool Command Processor script which removes all or certain metadata.
+ You can also use Attributes to record any kind of information about your files which should not go into potentially public metadata.
+ ...

Ain't it awesome