Emotion Filtering at the Edge
Emotion Filtering at the Edge
复制标题
边缘情绪过滤
DOI：
--
复制
发表时间：
2019
期刊：
影响因子：
--
通讯作者：
David Boyle
中科院分区：
文献类型：
--
作者：
Ranya Aloufi;H. Haddadi;David Boyle
语音控制设备和服务在消费物联网中变得非常流行。基于云的语音分析服务使用语音识别技术从语音输入中提取信息。因此，服务提供商可以建立非常准确的用户人口统计类别、个人偏好、情绪状态等档案，因此可能会严重损害他们的隐私。为了解决这个问题，我们在用户和云服务之间开发了一个保护隐私的中间层，以直接在边缘设备上清理语音输入。我们使用基于 CycleGAN 的语音转换从原始语音输入信号中删除敏感信息，然后重新生成中和信号进行转发。我们使用相对便宜的 Raspberry Pi 4 来实现和评估我们的情绪过滤方法，并表明性能准确性在边缘不会受到影响。边缘生成的信号与基于云的语音识别方法仅略有不同 (~0.16%)。对生成信号的实验评估表明，对说话者情绪状态的识别可以减少约 91%。
Voice controlled devices and services have become very popular in the consumer IoT. Cloud-based speech analysis services extract information from voice inputs using speech recognition techniques. Services providers can thus build very accurate profiles of users' demographic categories, personal preferences, emotional states, etc., and may therefore significantly compromise their privacy. To address this problem, we have developed a privacy-preserving intermediate layer between users and cloud services to sanitize voice input directly at edge devices. We use CycleGAN-based speech conversion to remove sensitive information from raw voice input signals before regenerating neutralized signals for forwarding. We implement and evaluate our emotion filtering approach using a relatively cheap Raspberry Pi 4, and show that performance accuracy is not compromised at the edge. Signals generated at the edge are shown to differ only slightly (~0.16%) from cloud-based approaches for speech recognition. Experimental evaluation of generated signals show that identification of the emotional state of a speaker can be reduced by ~91%.