Asset Bank Auto Tagging uses AI technology to automatically add keyword tags to image assets. There are four types of Auto Tags currently available:
- Object tags - Identifies objects in images based on shape, size, colour.
- OCR tags - Identifies any text in images.
- Face tags - Identifies whether an image contains a face or not.
- Landmark tags - Identifies famous landmarks in images.
By default, there is a limit of 10,000 Auto Tagging requests per month. If you have all four types configured, each asset you upload will make four requests. If you find that you regularly exceed your monthly request limit please contact our Customer Support Team.
You can change the tolerance of the Auto Tagging service in Admin > System > Auto Tagging Settings with a number between 0-100 which reflects the accuracy of the tags. The more accurate the service, the fewer tags you will generate. This has no impact on your monthly allowance.
Configuring Auto Tagging
Auto Tagging works by using Embedded Metadata Mappings to extract data from the image and insert it into an attribute. Below are the steps required for the three types of Auto Tagging.
Auto Tagging - Object Tags
In the Admin > Attributes area, add a new attribute for your object tags. In this example we have called it "Auto Tags (Objects)". We recommend that the attribute is a 'Text field (long)' attribute with keyword search links enabled. You can also configure this as a 'Keyword' attribute but this may conflict with your existing keyword list if you have one.
Ensure this attribute is visible upon upload and across any other areas of Asset Bank that you consider necessary in the Admin > Attributes > Display Attributes area. If you have the Asset Types feature enabled, remember to also select this new attribute for use with any applicable Asset Types.
In the Admin > Attributes > Embedded Metadata Mappings area, add a new mapping type as shown below.
Auto Tagging - OCR Tags
In the Admin > Attributes area, add a new 'Text Area (Long) attribute. You may wish to have keyword search links enabled. In this example, we have called it "Auto Tags (OCR)".
As above, make sure that the attribute is enabled for all applicable Asset Types and is displayed in all appropriate areas of Asset Bank by configuring the Admin > Attributes > Display Attributes area.
In the Admin > Attributes > Embedded Metadata Mappings area, add a new mapping type as shown below.
Auto Tagging - Face
In the Admin > Attributes area, add a new 'Dropdown' attribute. In this example, we have called it "Auto Tags (Face)".
On the Admin > Attributes page, click '[edit list values]' next to your newly created attribute. Here you will need to add a "No" value without an embedded data mapping value and a "Yes" value with the embedded data mapping value set to "face|faces".
Back in the Admin > Attributes area, edit your attribute and set the default value to "No"
As above, make sure that the attribute is enabled for all applicable Asset Types and is displayed in all appropriate areas of Asset Bank by configuring the Admin > Attributes > Display Attributes area.
In the Admin > Attributes > Embedded Metadata Mappings area, add a new mapping type as shown below.
Auto Tagging - Landmarks
In the Admin > Attributes area, add a new 'Text Area (Long) attribute. You may wish to have keyword search links enabled. In this example, we have called it "Auto Tags (Landmarks)".
As above, make sure that the attribute is enabled for all applicable Asset Types and is displayed in all appropriate areas of Asset Bank by configuring the Admin > Attributes > Display Attributes area.
In the Admin > Attributes > Embedded Metadata Mappings area, add a new mapping type as shown.
Testing and Using Auto Tagging
Once configured, you'll want to upload some test assets and check that it's working as expected. Below is an example of an asset with objects, text and a face in it.
Comments
0 comments
Please sign in to leave a comment.