平成26（2014）年度　一般研究２実施報告書

課題番号	26−共研−2023		分野分類			統計数理研究所内分野分類			c
						主要研究分野分類			2
研究課題名	Music emotion recognition based on Gaussian Process models
フリガナ 代表者氏名	マルコフ コンスタンティン ペトロフ Markov Konstantin Petrov				ローマ字		Markov Konstantin Petrov
所属機関	会津大学
所属部局	情報システム学部門
職　　名	准教授
配分経費	研究費	40千円		旅　費		34千円		研究参加者数		3　人

研究目的と成果（経過）の概要

Gaussian Processes (GPs) are becoming more and more
popular in the Machine Learning community for their ability
to learn highly non-linear mappings between two continuous
data spaces. Previously, we have successfully applied GPs
for static music emotion recognition. Dynamic or contin-
uous emotion estimation is more difficult task and there are
several approaches to solve it. The simplest one is to assume
that for a relatively short period of time emotion is constant
and apply static emotion recognition methods. A better ap-
proach is to consider emotion trajectory as a time varying
process and try to track it or use time series modeling tech-
niques. Some authors use Kalman filters to model emotion
evolution in time for each of four data partitions. For eval-
uation, KL divergence between the predicted and reference
A-V points distributions is measured assuming perfect test
samples partitioning. Our approach is similar since we also
use data partitioning, however, we apply model selection
method. In addition, we present novel dynamic music emo-
tion model based on GPs.
We invewstigated two state-space model based dynamic music
emotion recognition systems - one linear and one based on
Gaussian Processes. The preliminary results did not show
clear advantage of any system or feature set. This is proba-
bly due to the small size of the validation set. More detailed
experiments involving more data are planned for the future.

当該研究に関する情報源（論文発表、学会発表、プレプリント、ホームページ等）

1. Main research and experimental findings are summarized in the following
paper presentation:
K. Markov, T. Matsui, Dynamic music emotion recognition using State-Space
models, Working Notes Proceedings of the MediaEval 2014 Workshop, Barcelona, Catalunya, Spain, October 16-17, 2014.

2. Some of the results and methods are also covered in the following
journal paper:
K.Markov, T.Matsui, Music Genre and Emotion Recognition Using Gaussian Processes, Access, IEEE, v.2, pp.688-697, 2014.

研究会を開催した場合は、テーマ・日時・場所・参加者数を記入してください。

Research meeting has not been held.

研究参加者一覧
氏名	所属機関
Vazhenina Daria	会津大学
松井 知子	統計数理研究所