Service uses machine learning to enable film and TV locations to be found using photo rather than keyword searches


The first ‘visual search engine’ for TV and film locations, has launched. It’s AI-driven and uses machine learning to enable potential filming locations to be found using a photo rather than keywords.

It searches for locations with similar buildings, landscapes, and other features from a database of more than 500,000 images from 30,000 locations across the UK.

The service has been created by film location finder FilmFixer and image management software Image Data Systems and promises to “keep the nation’s location portfolio fresh, film friendly and accessible as never before.” is working with location agencies to add them to its network and further extend its coverage.

Click below for a demo of the visual search functionality.

Andrew Pavord, founder and chair of FilmFixer, said: ”Existing location websites rely on keywords. This is inherently flawed in a visual business such as ours. The brief location managers receive is usually a photo, accompanied by the request, ‘I need something that looks like this, but in Suffolk. It makes far more sense to drop that image into a search engine and add parameters such as, ‘inside the M25’ or ‘near Ipswich’, than to endlessly type in keywords to try and get a match.”