deepsing v0.1

seeing music though deeplearning videos stories art

deepsing translates music into unique, machine generated images keeping the sentiment and the arroused emotions - create visual stories using audio as input

Explore Demos
HOW & WHY

Built with state of the art deep learning

deepsing was born to materialize our idea of translating audio to images inspired by Futurama Holophoner. It works by performing attributed-based music-to-image translation and synthesizes visual stories according to the sentiment expressed by songs. The sentiment-aware generated images aim to induce the same feelings to the viewers, as the original song does, reinforcing the primary aim of music, i.e., communicating feelings.

  • 50000000
    Synapses
  • 1000
    Classes
  • 115200
    Unique Images

  • Input Signal

    Classify music segments based on valence and arousal

  • Audio to image

    Map audio and sentiment features to image categories

  • Stylize

    Enchance images' sentiment using neural style transfer

  • Generate content

    Deliver unique visual stories using Generative Adversarial Networks

Local Stories

Latest stories created

powered by imagination

Nikolaos Passalis

Postdoctoral Researcher

Nikos is a postdoctoral researcher in the Department of Informatics of Aristotle University of Thessaloniki. His research interests lie in the broad field of Deep Learning and Aritificial Intelligence.

Stavros Doropoulos

Software Engineer

Stavros is a Software Engineer with a strong passion for creating products using Machine Learning research, co-organizer of Thessaloniki Machine Learning meetup and currently CIO at DataScouting.

hack deepsing with us

we ❤ opensource
Github

Stay Tuned

Contact us for more information.