Emotion in Music task at MediaEval 2015
Journal
CEUR Workshop Proceedings
Journal Volume
1436
Date Issued
2015-01-01
Author(s)
Abstract
The Emotion in Music task is held for the third consecutive year at the MediaEval benchmarking campaign. The unceasing interest towards the task shows that the music emotion recognition (MER) problem is truly important to the community, and there is a lot remaining to be discovered about it. Automatic MER methods could greatly improve the accessibility of music collections by providing quick and standardized means of music categorization and indexing. In the Emotion in Music task we provide a benchmark for automatic MER methods. This year, we concentrated on a single task, which proved to be the most challenging in the previous years: dynamic emotion characterization. We put special emphasis on providing high-quality ground truth data and maximizing inter-annotator agreement. As a consequence of meeting a higher quality demand, the dataset both for training and evaluation is smaller than in the previous years. The dataset consists of music licensed under Creative Commons from the Free Music Archive, medleyDB dataset and Jamendo. This paper describes the dataset collection, annotations, and evaluation criteria of the task.
Type
conference paper