Abstract:Music emotion analytic is useful for many human-centric applications such as medical intervention. Existing studies have shown that music is a low risk, adjunctive and therapeutic medical intervention. However, there is little existing research about the types of music with identified emotions that have the most effect for different medical applications. We would like to discover various music emotions through machine learning analytic so as to identify modelsof how music conveys emotion features, and determine its effectiveness for medical intervention and treatment. We are developing a Human-centric Music Medical Therapy Exploration System which could recognize music emotion features from Chinese Folk Music Library, and recommend suitable music to playback for medical intervention and treatment. Our networked system is based on Support Vector Machine(SVM) algorithm to construct the models for music emotion recognition and information retrieval. Our main contributions are as follows: Firstly, we built up the Chinese folk music emotions and features library; secondly, we conducted evaluation and comparison with other algorithms such as Back Propagation(BP) and Linear Regression to set up the model construction for music emotion recognition and proved that SVM has the best performance; lastly, we integrated blood pressure and heartbeat data analytic into our system to visualize the emotion fluctuation in different music affection and make recommendation for suitable humancentric music medical therapy for hypertensive patients.

A Music Retrieval System Using Melody and Lyric

A MUSIC QUERY SYSTEM BY AUDIO RETRIEVAL

A Novel Approach Based On Fault Tolerance And Recursive Segmentation To Query By Humming

Human-centric Music Medical Therapy Exploration System

Query by Humming Systems Using Melody Matching Model Based on the Genetic Algorithm.

An efficient approach to humming transcription for query-by-humming system

Improving searching speed and accuracy of query by humming system based on three methods: feature fusion, candidates set reduction and multiple similarity measurement rescoring

K-BOX: a query-by-singing based music retrieval system.

Making Searchable Melodies: Human versus Machine.

A Noise Robust Content-Based Music Retrieval System for Mobile Devices

SENTENCE-BASED HUMMING RETRIEVAL

A Novel MIR System Based on Improved Melody Contour Definition

A New Spectral-Based Approach to Query-By-Humming for Mp3 Songs Database

Rapid Retrieval Technology of Query by Humming Based on Sentence Features

A Query by Humming System Based on Score

A Phrase-Level Piecewise Linear Scaling Algorithm for Melody Match in Query-by-Humming Systems

Efficient Vocal Melody Extraction from Polyphonic Music Signals

A Novel MIR Framework and Application with Automatic Voice Processing, Database Construction and Fuzzy Matching

Main Melody Extraction Using the Auditory Scene Analysis for the Humming Music Retrieval

A Music Retrieval Research and Implementation of New Methods

Content-Based Retrieval of Polyphonic Music Objects Using Pitch Contour