Page History

Introducing AI Tagging

Artificial Intelligence tagging is achieved in Chorus by integrating powerful third-party artificial intelligence services through our API. We create a separate metadata field for the tags that are added this way so that you can separate them from the ones entered by your users.

You can choose to scan your entire site or part of it by excluding specific folders or spaces. You can also edit these tags, as each picture will only be scanned once (unless a re-scan is forced). This can be particularly useful as AI tags can sometimes prioritize quantity over quality and are really useful when considered an augmentation of human capabilities rather than a replacement.

We offer three different kinds of metadata powered by artificial intelligence:

• Content recognition: keywords describing the content of a picture. This can range from very descriptive (“people”) to more contextual (“work meeting”)
• Landmark recognition: keywords naming famous landmarks shown in a picture. Examples could be the Tour Eiffel, or a famous mountain range.
• Optical Character Recognition (OCR): a text field containing all the recognizable words that can be found in a picture.

Setting up the Computer Vision Module

The Customer Success team at Third Light will set up the module for you, however it can be useful to know how your Chorus site's metadata configuration supports the AI tags, as follows.

The module requires three things to be created in your Chorus site:

• The corresponding metadata field for the service you want to activate, such as AI tags or OCR
• A Yes/No field that identifies the pictures that were already scanned, usually named “[…] Scanned”. You may need multiple ones if multiple services are activated.
• A User that acts as the service account for the AI scanning. This user is called “AI”.

Once these things are in place, Third Light periodically scans files that the AI user can see and edit. The AI user will also only scan files where the “Scanned” field is set to “No”.

Once a picture has been scanned, the “Scanned” field is automatically set to “Yes” to prevent re-scanning.

Restricting the scope of the Computer Vision Module

As part of the initial setup, the AI user and associated metadata fields are added to all the Spaces you want to scan.

Panel and fields will be created at site level, so that they can be easily added to any space.

If a folder that needs to be excluded from the scan, the folder's metadata field "Scanned" can be set to “Yes”. To control the scope of the Computer Vision module, the AI user can also be added or removed from Spaces as required.

If you add a new space to Chorus and want it scanned you can simply add the “AI tags” metadata panel to it, set the “Scanned” field to “No” and then add the “AI" user to the Space, in a role that can view and edit.

We recommend creating a role specifically for this user with just these permissions, as it reduces the risk to accidentally removing these permissions later.

This setup will be enough for the new space to be scanned.

If you want to only scan a folder in a Space, set the “Scanned” metadata field to “No” for that folder.

If you want to rescan a single picture, simply change the “Scanned” field to “No” for that picture.

You are here: