Berlin-Story-Booth

WHAT... if your computer listens to you when you tell a story and automatically generates a suitable background image?

Problem

The standard tool for supporting stories or presentations is PowerPoint. PowerPoint karaoke has even become a common term when a narrator refers to other people's slides and tries to interpret what they are supposed to say. But it should be the other way around. The visual support should adapt perfectly to the story being told.

Solution

The Berlin-Story-Booth uses the computer's microphone to listen to the narrated story. The spoken text is split into text fragments by a neural network and relevant keywords are identified. Then a suitable photo is selected from a picture library and displayed on the screen.

Challenges

The challenge was to ensure that the service was able to quickly and optimally understand the spoken word and display a suitable image without delay. We also needed to provide a well-maintained image library that did not contain inappropriate or poor-quality images.

Screenshots