MM supports a variety of different AI features.
Many of these settings need to be configured before they can be used.

Prerequisites

Required Roles

To use the AI features, users will require the role Ai_Add

Image requirements

By default, the rendition format “Large JPG Thumbnail” is used when generating AI content with Cognitive service.

If another rendition format is needed, it must adhere to the following requirements:

• The image must be presented in JPEG, PNG, GIF, BMP, WEBP, ICO, TIFF, or MPO format

• The file size of the image must be less than 20 megabytes (MB)

• The dimensions of the image must be greater than 50 x 50 pixels and less than 16,000 x 16,000 pixels

For guaranteed up-to-date requirements, we refer to Azure: https://learn.microsoft.com/en-us/azure/ai-services/computer-vision/overview-image-analysis?tabs=4-0

Enable AI settings

To enable AI settings, navigate to the AI configuration section

1. Click on Settings

2. Select “Portals”

3. Open the portal you wish to configure

4. Click Configuration

Open

Metadata fields

Each AI feature has to be coupled with a metadata field in your system.
For example, when you want AI to generate a description, you must define which note field the AI-generated description should be added to.

AI Description

A short description. Generated based on the content of the image.

Asset Types: Images
Metadata Type: note

AI Facial Recognition

Detects one or more faces in a picture, which can then be tagged as a specific person.

If trained on a face, it will recognize the person and auto-tag them in most cases.

Asset Types: Images
Metadata Type: editMultiComboValue

AI Tagging

Generates tags that describe what is on the image.

Asset Types: Images
Metadata Type: editMultiComboValue
Examples: building, outdoor, window, clouds

AI Transcript

Generates an audio transcript of a video.

Asset Types: Videos
Metadata Type: note

Explicit AI content detection

Detects whether there is any explicit content on the image.

Asset Types: Images
Metadata Type: multiComboValue
Possible Values: IsAdult, Is Racy, Is Not Explicit

Landmark AI detection

Detects a landmark on an image.

Asset Types: Images
Metadata Type: editMultiComboValue

Using the AI

When the AI settings are configured, they are ready to be used.
To use the AI, click on an asset to open the metadata editor.
Next, find the metadata fields you have configured and click the AI button below the metadata value field.
This will overwrite any existing metadata for the asset with the generated data.

Open

To auto-tag multiple image assets, select them and go into multi-edit mode.
Press the AI button, the same as you would when single-editing them, and wait for a couple of seconds.
If you want to compare the data, you can press the eye icon on the field, which will add the data of that metadata field to the asset selection side-bar on the right-hand side