Jump to content

Search the Community

Showing results for tags 'cloud vision'.

  • Search By Tags

    Type tags separated by commas.
  • Search By Author

Content Type


  • General
    • Announcements
    • Knowledge Base
  • Digital Asset Management
    • Daminion
    • Daminion Server
    • Web-Client for Daminion Server

Find results in...

Find results that contain...

Date Created

  • Start


Last Updated

  • Start


Filter by number of...


  • Start



Contact email



Found 1 result

  1. After accidentally or purposely deleting one or more AI Labels generated via the Google Cloud Vision API, there seems to be no way to recover them. Apparently, the AI Labels are not written to metadata via Sync or Write Tags to File and therefore cannot be restored by reading the metadata from the image file. Performing Actions -> Generate AI Labels again does not do anything (which is good, since it might incur additional fees by Google). There are (at least) two ways to delete AI Labels: Performing an Item -> Actions -> Read Tags from file deletes all existing AI Labels One or more AI Labels can be deleted manually in the Properties panel. If you do so accidentally, you may undo the action before saving the changes, but after that, they are gone for ever. The Tags Tree panel also shows a Delete option. I did not try it yet, but I assume it would work too. ... ? (I bet, there is more 😉 ) It seems to me, the values of all labels requested via the Google Cloudvision API still exist in the cloudvision table of the PostgreSQL database, but so far I could not find any way to restore or even visualize them in the Daminion client. I would greatly appreciate, if such a capability could be added to Daminion. Even though the Properties panel shows the words Add Tag ... under AI Labels, you cannot click it to manually add an AI Label (which doubtfully may make sense). However, if you select multiple image items and right click an AI Label which does not exist for all selected items, you may choose Assign 'xxx' to 'y' items. This could be a way to recover a deleted label, but only if you remember what it was and if another media item tagged with it, still exists. Thanks and a happy new year to everybody Wilfried
  • Create New...