The reviewed record of science sign in
Pith

arxiv: 1909.08500 · v1 · pith:EAAIRZZZ · submitted 2019-09-18 · eess.AS · cs.CR· cs.HC· cs.SD

Emotion Filtering at the Edge

Reviewed by Pith T0 review T1 audit T2 compute T3 formal T4 kernel pith:EAAIRZZZrecord.jsonopen to challenge →

classification eess.AS cs.CRcs.HCcs.SD
keywords edgeservicessignalsspeechvoicecloud-baseddevicesemotion
0
0 comments X
read the original abstract

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. In fact, signals generated at the edge 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%.

This paper has not been read by Pith yet.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.