Nowadays, music recommenders can learn users‘ personal preferences and consider contextual factors such as time and location. Our vision is to leverage latest improvements in automatic emotion measurement for context-aware music recommendation. As an initial step, we developed a conceptual framework with a first technical implementation – the EmpathicMusic app. This smartphone app is designed for application in automotive context and combines automatic facial emotion recognition with personalized music from Spotify. Our framework allows fast application of different music-emotion mappings and collection of explicit and implicit user feedback. In future work, we aim to provide more insights into the complex relationship of emotional states and music recommendation during real-life usage. Below are two demo videos: 1.) showing the start-up of our EmpathicMusic app including the Spotify app as well as further functionalites of our app (the baseline mood rating, song skipping, artist selection, view for front-camera positioning) and 2.) showing the ending of a journey including a second baseline mood rating, rating of the suggested music and the exemplary start of a one-time demographics and personality questionnaire.

© Jonas Köckerling
Nach oben scrollen