Commentary

Royal Caribbean AI Tool Sets Vacation Photos To Music

Artificial intelligence has been finding its way into countless applications, and now even the creation of personalized music soundtracks matched to vacation photos.

A new system called SoundSeeker is designed to use machine learning to create original soundtracks based on the content of a few photos submitted by an individual.

Royal Caribbean teamed with experts from Berklee College of Music in Boston to create Sound Seeker, a unique song generator. The team created a roadmap of tools that help determine the musical pairing to photos, accounting for tempo, pitch and instruments.

To use SoundSeeker, which is free, a person uploads three photos at the Royal Caribbean SoundSeeker site, which can be connected via Facebook or Instagram for later sharing, or photos can be uploaded without any social media connections and then emailed to yourself or a friend.

advertisement

advertisement

SoundSeeker uses machine learning based on more than 600 hours in which a team of musicians and technologists reviewed hundreds of music tracks along with 10,000 photos, matching each of the 3 million combinations to one of 10 moods.

The AI in SoundSeeker then uses Google Cloud Vision to identify objects, facial expressions and colors in a person’s photo by referencing the roadmap developed by the music theory experts at Berklee. SoundSeeker then determines the music elements that correspond to each mood in the photo to compose an audio and video photo album.

More than 1 million unique tracks, created exclusively for Royal Caribbean, can be generated. They include 90s hip-hop, rock, modern and electronic dance music.

The music videos created are very quick, since just three photos have to be identified, and the creation of a video is complete within seconds. In this first iteration of SoundSeeker, the videos produced are quite short, but the system at least shows where this is heading.

 

Next story loading loading..