Object ID is an advanced Visual AI feature within PhotoShelter for Brands. It allows organizations to tag specific items (such as products, props, landscape features, birds, mascots, etc) within their image library and automatically surface other instances of those items.
Unlike standard keyword tagging, Object ID uses visual recognition. It does not require you to upload a separate set of "training images." Instead, it learns directly from your Library: as you identify objects in your existing photos, the system learns to recognize them in the rest of your archive.
Setup
Your Customer Success Manager will be your point of contact for all things ObjectID and will guide you through the entire setup process.
Who Can Use This Feature?
- Library Staff: Administrators and Editors can identify new objects and confirm AI suggestions.
-
Invited Users: Once an object is tagged and confirmed, it becomes searchable for all users with access to those assets.
How Object ID Works
Object ID operates on a "learn-as-you-go" model. The process begins when you manually identify an object in one of your photos. The AI then analyzes the visual characteristics of that object and scans your library to find matches.
The system improves over time: The more instances you confirm or remove, the smarter and more accurate the detection becomes.
ObjectID can be applied at a per-Collection level:
The Global "All New Uploads" ObjectID setting is found under your Automated Metadata Settings page:
https://library.photoshelter.com/settings/metadata/automated-metadata-settings
ObjectID Filtering
ObjectID can be told NOT to tag specific words/tags as well. This setting can be found in your Library under Settings > Metadata > Object ID: