Embedding Image Search
Category: Technology
An image search that translates images and queries into numerical vectors (embeddings) and finds results by content similarity – not just by keywords.
In embedding image search, image content and queries are converted by an AI model into high-dimensional vectors. Content-similar subjects lie close together in this vector space, so the search works by meaning rather than exact keywords.
In the portal this enables a visual search: you find matching subjects via a description or an example image, even when the stored keywords do not match literally.
Find images matching this topic via visual search in the portal.
Browse all images →Explained by Daniel Richter · Expert in gigapixel photography · Gigapixel GmbH