Case Study: Adobe achieves advanced image recognition and search with Elastic

Image recognition and search at Adobe with Elasticsearch and Sensei

Adobe needed to support powerful image and content search across products like Adobe Stock, Lightroom, and its cloud suites while handling massive, mostly non-textual data at scale. The company used Elastic’s Elasticsearch and self-managed Elastic Stack, alongside Adobe Sensei, to power search for more than 10 billion documents and over 130 million stock images.

Elastic helped Adobe implement custom Elasticsearch plugins for similarity and ranking, enabling features like image recognition, face and object detection, auto-tagging, and image-similarity search. The result was a highly responsive search experience backed by 18 production clusters and a live ingestion rate of about 6,000 documents per second, with Adobe also migrating nearly 3.5 billion documents from Amazon’s Elasticsearch Service to its own Elastic-managed environment.

Adobe

Baldo Faieta

Senior Computer Scientist