Image Similarity Search

We enable you to integrate image similarity search capabilities into your products and applications. Our similarity API is using deep learning to provide you the best results.

How does it work?

  • Upload images which should be part of the search results (please see supported sources)
  • Build the index and wait till it's done
  • Send requests against our API endpoint to get similarity search results for your uploaded images

Supported Sources

Google Cloud Storage

We support Google Cloud Storage, you can either import one or multiple Google Bucket folder or multiple single bucket paths. In order to allow us to access your files inside of Google Cloud Storage please add the following service account as member of your bucket service@io-annotator.iam.gserviceaccount.com. The service account requires the following rules in order to list and read the files in your bucket

  • Storage Object Viewer
  • Storage Legacy Bucket Reader

http and https

Images which are public accessible via http or https can be used

Amazon S3

Amazon S3 support is on the roadmap, if you want to use our Similarity Search with Amazon S3 please let us know.

Azure Storage

Please contact us if required.

Supported Filetypes

  • png
  • jpg
  • gif (not animated)
  • gif (animated, will take first frame)
  • bmp
  • tiff

If you need other file types please let us.

Index Rebuild

If you add new images to the index an index rebuild is required. There is no downtime during index rebuild. We support two different types of index rebuilds:

  • Manual Index Rebuild
    Can be triggered via button inside of the UI

  • Scheduled Index Rebuild
    Can be configured as scheduled job, for example: run every day at midnight

  • API Index Rebuild (soon)
    Can be triggered via API call