Scholarlink

Music signals and electrocardiogram (ECG) signals have different semantics. Quantifying their relationships has a wide range of real-world applications such as personalized music recommendation, music composition, and music digital therapeutics. However, this task remains underexplored due to the absence of an explicit and measurable link between music and ECG signals. In this paper, we propose CardioMusic to bridge this gap based on a shared valence-arousal (VA) emotional space. CardioMusic is a two-branch deep learning model which first embeds music signals and ECG signals into latent features, then projects the separate latent features into a shared emotional space by learning their emotional similarities using VA values. Experimental results demonstrate that CardioMusic effectively captures the VA values of both music and ECG signals and aligns them within the shared VA emotional space.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Coordinating Music Signals and Electrocardiogram Signals in a Shared Valence-Arousal Emotional Space

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Jiaze Wang,
  • Shijia Geng,
  • Zizheng Chen,
  • Yuhu Han,
  • Deyun Zhang,
  • Weilun Xu,
  • Lixin Wang,
  • Yanqing Zhao,
  • Shenda Hong,
  • Yuxi Zhou

摘要

Music signals and electrocardiogram (ECG) signals have different semantics. Quantifying their relationships has a wide range of real-world applications such as personalized music recommendation, music composition, and music digital therapeutics. However, this task remains underexplored due to the absence of an explicit and measurable link between music and ECG signals. In this paper, we propose CardioMusic to bridge this gap based on a shared valence-arousal (VA) emotional space. CardioMusic is a two-branch deep learning model which first embeds music signals and ECG signals into latent features, then projects the separate latent features into a shared emotional space by learning their emotional similarities using VA values. Experimental results demonstrate that CardioMusic effectively captures the VA values of both music and ECG signals and aligns them within the shared VA emotional space.