人工智能培训

搜索

人工智能论文:音乐流派分类的自下而上广播神经网络(Bottom-up Broadcast Neural Network For Music Genre Clas

[复制链接]
mxs810 发表于 2019-1-28 12:07:46 | 显示全部楼层 |阅读模式
mxs810 2019-1-28 12:07:46 110 0 显示全部楼层
人工智能论文:音乐流派分类的自下而上广播神经网络(Bottom-up Broadcast Neural Network For Music Genre Classification)基于视觉表现的音乐流派识别在过去几年已成功探索。最近,人们越来越关注尝试卷积神经网络(CNN)来完成任务。然而,大多数现有方法采用图像识别中提出的成熟CNN结构而没有任何修改,这导致学习特征不足以用于音乐类型分类。面对这一问题的挑战,我们充分利用了音频谱图中的低级信息,并在本文中开发了一种新颖的CNN架构。提议的CNN架构将长的上下文信息考虑在内,从而为决策层传递更合适的信息。几个基准数据集的Variouse实验,包括GTZAN,Ballroom和Extended Ballroom,已经验证了所提出的神经网络的出色表现。代码和型号将在“ttps://github.com/CaifengLiu/music-genre-classification”中提供。
Music genre recognition based on visual representation has been successfullyexplored over the last years.Recently, there has been increasing interest inattempting convolutional neural networks (CNNs) to achieve the task.However,most of existing methods employ the mature CNN structures proposed in imagerecognition without any modification, which results in the learning featuresthat are not adequate for music genre classification.Faced with the challengeof this issue, we fully exploit the low-level information from spectrograms ofaudios and develop a novel CNN architecture in this paper.The proposed CNNarchitecture takes the long contextual information into considerations, whichtransfers more suitable information for the decision-making layer.Variousexperiments on several benchmark datasets, including GTZAN, Ballroom, andExtended Ballroom, have verified the excellent performances of the proposedneural network.Codes and model will be available at"ttps://github.com/CaifengLiu/music-genre-classification".人工智能论文:音乐流派分类的自下而上广播神经网络(Bottom-up Broadcast Neural Network For Music Genre Classification) RmwA3jFwUp5X3syF.jpg
URL地址:https://arxiv.org/abs/1901.08928     ----pdf下载地址:https://arxiv.org/pdf/1901.08928    ----人工智能论文:音乐流派分类的自下而上广播神经网络(Bottom-up Broadcast Neural Network For Music Genre Classification)
回复

使用道具 举报

您需要登录后才可以回帖 登录 | 立即注册

本版积分规则 返回列表 发新帖

mxs810当前离线
新手上路

查看:110 | 回复:0

快速回复 返回顶部 返回列表