Skip to content

A Neural Network for Explaining mid-level features and using them for Music Emotion Recognition

Notifications You must be signed in to change notification settings

Jarvis1000x/Music_Emotion_Recognition

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 
 
 
 
 

Repository files navigation

Explainable Music Emotion Recognition using Mid-Level features

This Repository is an implementation of Towards Explainable Music Emotion Recognition The Route via Mid-level features by S. Chowdhury et al. The model tries to give a musically meaningful and intuitive explanation for its Music Emotion predictions, a VGG-style deep neural network has been used that learns to predict emotional characteristics of a musical piece together with (and based on) human-interpretable, mid-level perceptual features.

Dataset

For datasets, the Aljanaki & Soleymani’s Mid-level Perceptual Features dataset provides mid-level feature annotations. For the actual emotion prediction experiments, the Soundtracks dataset has been used, which contains the Aljanaki collection as a subset, and comes with numeric emotion ratings along 8 dimensions.

About

A Neural Network for Explaining mid-level features and using them for Music Emotion Recognition

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages