How can I build my own image database?

Nowadays, a professional image database is a necessary purchase not only for photographers and media companies. In more and more companies, organizations and institutions, marketing, sales and public relations teams face the challenge that the number of digital images has grown exponentially over the past two decades. Often, image inventories quickly exceed several thousand, tens of thousands or even hundreds of thousands of images. Therefore, it makes absolute sense to deal intensively with the question of how to build up your own image or photo database in order to find the right motif quickly and easily in the future.

Five reasons to build your own image database:

1. Relief of employees (time, resources, legal requirements)
2. Facilitated compliance with legal requirements (e.g. copyrights, personal releases, data protection requirements)
3. Use synergy effects (e.g. no duplicate purchases of images)
4. Improved quality of public relations and marketing
5. Increase the return on investment (ROI) of digital images or media files

What is the difference between a folder-based filing structure and an image database?

Many organizations have filed and stored their images on classic media drives in a (sometimes complex) folder structure. When organizing images in a folder structure, however, you quickly reach the limits of what is possible when trying to assign images to multiple topics. For example, a picture of a sporting event might capture a player, a sponsor logo, and an emotion. If I want to add this information to the image, I have few options in a one-dimensional folder structure.

Option A: I create three folders (player, sponsor and emotion) and place a copy of the image in each folder.

Option B: I create a folder and keep the information in the folder or file name.

Neither option A nor option B are structurally sound alternatives, since you either create masses of duplicates or you end up with a confusing and poorly searchable file structure.

So what might an alternative look like in the form of a professional image database?

When setting up your own image database, it is important to first familiarize yourself with the technical possibilities of a professional image management. Unlike working with a one-dimensional folder structure, such a solution offers you numerous possibilities to link your images with supplementary information.

Examples of a multidimensional data structure of an image database:

• Tags / keywords
• Categorization by media type
• Organization in virtual albums / collections
• Usage rights / license information
• AI-based data analytics
  • number of people
  • personal identification
  • dominant accent colors
  • automated object recognition
  • mood / emotions

What are my requirements for my own image database?

The following questions can help you create your requirements profile for building your own image database for your company:

• Should only images or also other media files be processed?
  (e.g. videos, graphic or audio files, presentations, documents)
• What file formats are used?
  (e.g. JPEG, PNG, TIFF, WebP, AI, SVG, MP4, AVI, MOV, PDF, etc.)
• Which metadata schemas and formats should be used?
  (e.g. Exif, IPTC-IIM, XMP, Dublin Core)
• Who should work with the system internally in the future?
  (e.g. employees from Marketing and PR department, Corporate Communications, Human Resources, Sales).
• Which external stakeholders need access to the database?
  (e.g. agencies, customers, partners or press)
• Are legal issues relevant to license management, e.g. of stock images or otherwise purchased imagery?
• Is AI-based facial recognition needed for automatic identification of specific individuals?
• Are there company-specific processes or requirements that need special consideration?
• Do special data protection requirements (beyond the GDPR) have to be observed?
  (e.g., particularly high security requirements due to confidentiality obligations).

Modern image management platforms are usually cloud-based and feature a self-explanatory user interface as well as a high degree of flexibility and security.

Accelerated work through mass indexing, here with the cloud-based solution for image databases from teamnext

Once you have found the right solution for you, you should test it thoroughly yourself to get your own idea of the system’s structure and flexibility.

Before you start building your own image database, you should first plan the future data structure.

How do I find the right data structure for my company?

By working out your individual data structure, you lay the foundation for the long-term success of your own image database. Therefore, it is crucial in this phase to get as heterogeneous an opinion as possible from future users from different areas of your company. This not only validates the data structure, but also increases the acceptance of the internal database among your colleagues.

Depending on the content focus, the top-level data structure of your image database could look like this:

1. Chronological sorting (e.g. season / competition in professional sports)

1. • Bundesliga season 21/22
   • Cup games season 21/22
   • Friendlies season 21/22
   • etc.

2. Organizational areas (e.g. city administration)

1. • Unit I
   • Unit II
   • Unit III
   • etc.

3. Product or article groups

1. • Product group A
   • Product group B
   • Product group C
   • etc.

4. Content structure

1. • Press conferences
   • Products
   • Projects
   • Logos and templates
   • etc.

There is no right and wrong when creating the data structure, as long as the new structure fits your organization.

Once the filing structure for your own image database has been created, the next step is to build up the keyword catalog so that you can add supplementary information to your images.

How do I create an individual catalog of keywords or tags?

When creating a keyword catalog for your image database, which is also referred to as a thesaurus in technical language, the heterogeneity of the people involved is just as important as when developing the filing structure. Ideally, you should create a list and, as a first step, collect all the keywords that seem useful to help you find images in your daily work. In the second step you group or categorize the keywords. This gives you an overview of whether all areas of the organization have already been taken into account.

Tag management within teamnext’s professional image database software

How do I structure a keyword tree?

Try to work with generic generic terms as much as possible. The deeper a level, the more specific the terms should become. But don’t get lost in the nitty gritty. A structure depth of 2 to 4 is usually sufficient if you do not want to build up a scientific image database. In addition, synonyms should be created for some terms. Finally, an image tagged with “fall” should also be found when “autumn” is entered as a search term. A comprehensive synonym directory is therefore always part of a well-maintained keyword structure. Since with a good image database you can specify exactly which keywords are output in the frontend (synonyms are usually only indexed), this is also referred to as controlled vocabulary.

Example categories of a keyword directory:

Season

• • Spring
  • Summer
  • Fall (synonym: autumn)
  • Winter

Location

• • Germany
    • Bavaria
      • Augsburg
      • Munich
      • Nuremberg
      • etc.
    • Baden-Württemberg
    • Berlin
    • etc.
  • United States of America (synonym: USA)
    • Alabama
      • Montgomery
    • Alaska
    • Arizona
    • etc.

Which images should be imported into your own image database?

With the creation of an own image database, sooner or later the question arises whether all images from the previous stock should be imported into the new database or not. Many image and media collections have grown over many years, so a “digital spring cleaning” is a good idea in many organizations. In doing so, you should create a common understanding in advance of what characteristics distinguish “good” images, because only such images should be available in the database in the long term.

What are the characteristics of a “good” image?

• Content is clearly understandable
  • persons are well met
  • sufficient information content and significance
  • categorizable in: creative / informative / documentary
• Recognizable target group focus
• Good image composition (e.g. via golden section, symmetry, image depth, framing, leading lines, etc.).
• Contemporary or innovative imagery (images should not appear old-fashioned, often the case with older stock photos)
• Technical quality of the image
  • little image noise
  • no visible artifacts (e.g. due to too much compression)
  • clear focus (sharpness / blur)
  • balanced image exposure
  • usable absolute resolution (at least 1 MP)
  • high contrast range (dynamic range)
  • successful white balance / color fastness
  • not upscaled or interpolated too much

Test the image management software from teamnext

If you’ve read this far, you’re probably also interested in putting what you’ve learned into practice. At teamnext, we have developed the Media Hub, a cloud-based image management solution that not only meets the requirements discussed here, but is also easy to use and has AI-based features that will speed up your workflows enormously. If you just want to try our solution, you can start a free 14-day trial here. In addition, you can book an appointment for an online product demo with one of our experts at any time. Simply use our contact form for this purpose.

Everything in view – with the professional solution for image databases from teamnext