The same face across the whole archive

A person named once is a person found everywhere. Faces are located in images and in video, grouped, and matched against the people your curators have named, so a name given to one photograph reaches every other picture and every recording that person appears in. Each recognised face links to a person record carrying its appearances across the archive, and those records are curated in one place.

Faces detected and matched to person records; public-people packs, available in certain countries and regions, provide automatic name labelling, with aliases, of thousands of public figures. Where an installation’s licence covers a region, the public figures who appear in your material are labelled with their names and known aliases from the day they are ingested, without anyone naming them first.

Your own people stay your own. The people your team names are recognised only inside your archive. They never contribute to anyone else’s recognition, and a name your team has entered always takes precedence over one a pack supplies.

In video, at their own timestamps. Faces found in recordings are recorded under your installation’s own people, exactly like faces in images. In the viewer each appearance is outlined on the frame at its own moment, labelled with the person’s name.

And when two records turn out to be one person, the review shows the similarity percentage beside the pair, and each action states what it will do before anyone takes it.

Faces across media are shown working on the discovery page.