人工智能培训

搜索

深度学习论文:用于肺部呼吸相位检测的卷积神经网络(Convolutional neural network for breathing phase detect

[复制链接]
imissa 发表于 2019-3-26 11:52:02 | 显示全部楼层 |阅读模式
imissa 2019-3-26 11:52:02 469 0 显示全部楼层
深度学习论文:用于肺部呼吸相位检测的卷积神经网络(Convolutional neural network for breathing phase detection in lung  sounds)我们应用深度学习来创建肺部录音中的呼吸相位检测算法,并且我们比较了算法检测到的呼吸阶段,并由两位经验丰富的肺部声音研究人员手动注释。 Ouralgorithm使用卷积神经网络,其中光谱图作为特征,无需明确指定特征。我们使用比文献中先前所见更大的三个子集来训练和评估算法。我们使用twomethods评估了该方法的性能。首先,商定呼吸阶段的离散计数(在一对盒子之间使用50%重叠)显示出与肺部声音专家的平均一致性,即灵感为97%,呼气时为87%。第二,协议时间的分数(以秒为单位)给出吸气的伪卡伯值(0.73-0.88)高于呼气(0.63-0.84),显示平均灵敏度为97%,平均特异性为84%。利用这两种评估方法,注释器和算法之间的一致性显示了算法的人类水平表现。所开发的算法对于检测声音记录中的呼吸阶段是有效的。
We applied deep learning to create an algorithm for breathing phase detectionin lung sound recordings, and we compared the breathing phases detected by thealgorithm and manually annotated by two experienced lung sound researchers.Ouralgorithm uses a convolutional neural network with spectrograms as thefeatures, removing the need to specify features explicitly.We trained andevaluated the algorithm using three subsets that are larger than previouslyseen in the literature.We evaluated the performance of the method using twomethods.First, discrete count of agreed breathing phases (using 50% overlapbetween a pair of boxes), shows a mean agreement with lung sound experts of 97%for inspiration and 87% for expiration.Second, the fraction of time ofagreement (in seconds) gives higher pseudo-kappa values for inspiration(0.73-0.88) than expiration (0.63-0.84), showing an average sensitivity of 97%and an average specificity of 84%.With both evaluation methods, the agreementbetween the annotators and the algorithm shows human level performance for thealgorithm.The developed algorithm is valid for detecting breathing phases inlung sound recordings.深度学习论文:用于肺部呼吸相位检测的卷积神经网络(Convolutional neural network for breathing phase detection in lung  sounds) fY2HXhOQHa0u0Lyy.jpg
URL地址:https://arxiv.org/abs/1903.10251     ----pdf下载地址:https://arxiv.org/pdf/1903.10251    ----深度学习论文:用于肺部呼吸相位检测的卷积神经网络(Convolutional neural network for breathing phase detection in lung  sounds)
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

imissa当前离线
新手上路

查看:469 | 回复:0

快速回复 返回顶部 返回列表