Try the teamnext | Media Hub now for free
Try the Media Hub now
Test 14 days without obligation
Test phase ends automatically
YOU CAN REACH US AT
teamnext GmbH & Co. KG
34117 Kassel, Germany
+49 561 473 95 98 – 0
Table of contents
In this how-to post, you’ll learn why tagging images is so important and what makes for good keywording. However, to avoid misunderstandings, we would first like to briefly clarify the terminology.
The process we call tagging is also called subject indexing or keywording. Basically, this means the same thing: providing documents, media and other content with so-called descriptors: text elements that we know as keywords or catchwords. The purpose is clear: it is to improve the findability of content within an archive or database. Classical librarianship is familiar with the principle from subject catalogs (usually in the form of card indexes). But historical things should not concern us any further. We will limit ourselves here to digital media, focusing in particular on the tagging of photos.
The main reason why keywording is important has already been mentioned: it increases the findability – in our case of files within a digital database. Furthermore, a good keywording provides more structure and overview, since keywords are usually not created one-dimensionally, but subject to a multi-dimensional hierarchy. But more about that later. First, let’s take a step back and look at the search process.
Searching a digital photo collection using keywords is, of course, much more convenient than shuffling through numerous folders each time to find a desired image. Don’t take this the wrong way: A well thought-out folder structure can certainly be helpful when storing image collections, but once a collection has reached a certain size (around 10,000 images upwards), it is simply appropriate to invest time in systematic keywording. The emphasis here is on systematic, because you gain little to nothing from tagging only part of your images, after all, the other part will not be picked up in a keyword-based search. So don’t approach the keywording project half-heartedly. The motto here should be ” all or nothing“.
If you’ve been looking for this article, it’s probably also because you’re faced with the task of having to open up a larger stock of images yourself. You may even already have a program in mind that you would like to use for keywording. This is because on-board tools of the operating system are not sufficient here, even if Windows Explorer, for example, offers rudimentary keywording options for image formats such as JPEG or TIFF. After reading the following sections, you will certainly be in a better position to judge whether the software at your disposal meets the requirements for systematic indexing.
Of course, you could just go ahead and enter the terms that first come to mind when viewing an image. From painful experience I can say that there are many photographers who do it exactly the same way. Please do not. If you want to tag images professionally, you need a controlled vocabulary (also known as a thesaurus ). So you need to clarify which terms should become part of this vocabulary and which should not.
To create a controlled vocabulary, start by thinking about what subjects you photograph most often or what subjects are most often in your inventory. Depending on the orientation, the most important categories should then be found. Assume the stock contains mostly subjects attributed to nature photography. Then, for example, the following categorical classification might be useful:
One thing in advance: Categories are nothing else than the supreme concepts, and thus they are of course also supreme concepts. Now, in the next step, we want to find meaningful terms for the layers below. Within the category “Animals” these would be for example:
Subheadings to “mammals” would be:
On the lowest level would be the name of the specific object, so in our case the species name or for farm and domestic animals also the name of the breed. An exemplary hierarchy could therefore look like this:
Animals > Mammals > Dogs > Domestic dogs > Dachshund > Shorthaired Dachshund
The scheme can be applied to any subject. You start with general terms and get more specific level by level, e.g.:
Technics > Communication technology > Mobile telephony > Smartphones > iPhone > iPhone 14 Pro Max
Of course, these do not always have to be terms that refer to objects. Concepts expressed in images can also be keyworded. For example, think of emotions, moods, or styles. A corresponding example would then look structurally like this:
Concepts > Image Mood > Romantic
As you have seen, keywords are ideally created multidimensionally. This results in a structure familiar from biological systematics or phylogenetic trees. Therefore, one also speaks of a keyword tree. However, you should not create too many branches, otherwise you run the risk of getting lost in the minutiae. We recommend a maximum branching depth of 5 or 6.
You may now be wondering what the direct benefit of this tree structure is. The answer to this is simple and legitimizes the extra effort: you’ll have to type in far fewer terms by hand in the long run. Because with a good keywording solution, it is enough to enter a sub-term (or the lowest term), e.g. “shorthaired dachshund”, and all higher-level terms created in the structure will be added automatically. With a single input, the following keywords would then already be on the image:
Animals, mammals, dogs, domestic dogs, dachshund, shorthaired dachshund
But it gets even better. With a professional solution, it is always possible to store synonyms as well. For example, you could add the following terms to the “Mammals” entry:
For “dachshund”, for example, you could create the following synonyms:
Useful synonyms can also be found for the remaining generic and subordinate terms. Once the synonym directory is up and running, you can save a lot of time, because just typing “shorthaired dachshund” now leads to keywords like:
animal, animals, fauna, mammal, mammals, Mammalia, dog, dogs, Canidae, domestic dog, domestic dogs, Canis lupus familiaris, dachshund, dachshunds, shorthaired dachshunds
While the tree structure ensures that images are found even when searching with general terms or combinations of terms (e.g. with “animals + cute”), the synonym directory catches alternative spellings that would otherwise have led nowhere. Sometimes it can even be helpful to deliberately create incorrect spellings as synonyms. A typical example: What spelling is correct? pharoah or pharaoh? Well, it’s the latter. But many people spell it incorrect. So if the keywording is also intended for customers or other external users, it is generally advantageous if typical misspellings are also included in the synonym directory.
However, there are also tricky cases that cannot be absorbed one hundred percent even with the best planning. Here is an example:
When someone searches for “Eifel Tower”, you can almost be sure that the search is for the Eiffel Tower in Paris. So one might be tempted to put “Eifel Tower” as a synonym for “Eiffel Tower”. However, there is also an Eifel Tower, and it is in the district of Mayen-Koblenz in the Eifel. You can already see where this is going, and before we get further lost in details, let’s move on to the next question.
There are properties related to composition or format that can be queried for each image. Let’s go through them in order, starting with the format. Here you could enter as a keyword whether the image format is:
By the way, modern keywording tools add this information automatically.
Next, there are technical and creative means of various kinds. In the field of photography, for example, we work with the setting of the angle of view and focal length, accordingly there are categories like:
Additionally there are different perspectives that shape an image, e.g.:
In the case of images of individual objects or persons, there is a whole repertoire of terms in relation to the image detail, which we can only hint at here. Thus, however, you already know what to look for when it comes to practical implementation. These are terms like:
You can already see where the journey is headed. Of course, there are many more terms for image composition, image environment or photographic technique. A few more are mentioned so that you can better appreciate the scope that such a project can take on if you want to make it perfect. Frequently queried are characteristics such as:
In stock photography, information about the presence of people in the image is also provided in the form of keywords, e.g.:
Frequently, location and time information is also provided for keywords. This could be information such as:
In addition, there are numerous keywords that result from legal facts. For example, it is about the existence of model releases or questions about license models and permissible uses:
Finally, we would like to apply what has been presented and show an example of good keywording. Since this article does not deal with the creation of the image description (caption) or the image title, we will limit ourselves to the creation of the keywords. In order to create a better overview, we will group the terms and distinguish between content-related, conceptual and formal keywords.
Content keywords for the above image:
animal, animals, fauna, mammal, mammals, Mammalia, dog, dogs, Canidae, domestic dog, domestic dogs, pet, pets, Canis lupus familiaris, dachshund, dachshunds, shorthaired dachshunds, spotted, multicolored, male, males, scarf, stairs
cute, loyal, faithfulness, doggy eyes
Photography, color photo, landscape, portrait, side view, central perspective, camera view (eye contact), outdoor, brown, brown tones, nobody
If you’ve read this far, you’re probably interested in putting what you’ve learned into practice. Therefore, we would like to introduce you to our comparatively easy-to-use solution that combines two technologies: On the one hand, you can use our Media Hub to keyword images and videos in the classic way (image by image or via mass keywording), and on the other hand, the system is equipped with a module for automatic keywording. Using machine learning (this is also referred to as AI, or artificial intelligence), the module can not only identify everyday objects in images, it can also recognize specific people, specific products, and individual tastes and logos through targeted training and generate corresponding keywords. A central keyword management with synonym directory is of course also included.
If you would like to get to know our solution, then you can start a free 14-day trial here. In addition, you can of course book an appointment for an online product demo with one of our experts at any time. Simply use our contact form for this purpose.