Thesis

Github
Sonification: Extending the Visual Arts experience
My recent AI master thesis focused on neural networks and computer vision. Using neural networks such as Places and Soundnet and the computer vision library OpenCV a framework came to be that can transform paintings into sound Using multiple technologies paintings are automatically converted into sound. By extracting color with OpenCV the HSV values are linked to sound properties like chords and melodies. Using Places and Soundnet, sound of the scene the painting exhibits is added as a background sound. SuperCollider is responsible for synthesising the properties of the painting extracted by the models and algorithms.