论文部分内容阅读
为了能实时监测听众的情绪状态变化并据此调整音乐播放列表,本文中,我们基于便携式干电极脑电数据提出了一种脑电驱动的个性化情绪音乐推荐系统的算法框架,并在Android平台上进行了初步开发实现。我们以效价和唤醒度二维情绪模型为基准,将脑电和相应种子歌曲映射到各个情绪坐标象限内,从而建立映射关系。然后应用美尔频率倒谱系数分析音乐库中各歌曲与种子歌曲之间的相似度并进行排序。最后,在播放歌曲阶段,我们通过脑电来识别听众的情绪状态,根据事先获得的情绪状态匹配关系播放并实时调整相应的歌曲曲目列表。
In order to monitor the emotional state changes of listeners in real time and adjust the music playlist accordingly, in this paper, we propose an EEG-based algorithm framework of personalized emotion music recommendation system based on portable dry electrode EEG data, On the initial development to achieve. Based on the two-dimensional emotional model of potency and arousal, we mapped the EEG and the corresponding seed songs to each emotion coordinate quadrant to establish the mapping relationship. Then, the frequency of the ceilings in the music library is used to analyze the similarity between the songs in the music library and the seed songs by using the frequency of the celestial frequency. Finally, during the song playing stage, we recognize the emotional states of the listeners through EEG, and play and adjust the list of the corresponding song tracks in real time according to the pre-obtained emotional state matching relationship.