Personalised models for speech detection from body movements using transductive parameter transferReport as inadecuate




Personalised models for speech detection from body movements using transductive parameter transfer - Download this document for free, or read online. Document in PDF available to download.

Personal and Ubiquitous Computing

, Volume 21, Issue 4, pp 723–737

First Online: 16 February 2017Received: 22 July 2016Accepted: 01 February 2017

Abstract

We investigate the task of detecting speakers in crowded environments using a single body worn triaxial accelerometer. Detection of such behaviour is very challenging to model as people’s body movements during speech vary greatly. Similar to previous studies, by assuming that body movements are indicative of speech, we show experimentally, on a real-world dataset of 3 h including 18 people, that transductive parameter transfer learning Zen et al. in Proceedings of the 16th international conference on multimodal interaction. ACM, 2014 can better model individual differences in speaking behaviour, significantly improving on the state-of-the-art performance. We also discuss the challenges introduced by the in-the-wild nature of our dataset and experimentally show how they affect detection performance. We strengthen the need for an adaptive approach by comparing the speech detection problem to a more traditional activity i.e. walking. We provide an analysis of the transfer by considering different source sets which provides a deeper investigation of the nature of both speech and body movements, in the context of transfer learning.

KeywordsSocial signal processing Wearable sensors Social actions Transfer learning Human behaviour An extended abstract version of this paper is published in UBICOMP 2016 with the title of -Speaking Status Detection from Body Movements Using Transductive Parameter Transfer- 9. In addition to preliminary results presented in the UBICOMP paper, current paper presents an analysis of connection between speech and body movements, provides comparisons with the state-of-the-art methods and different implementations of TPT, analyses source quality in transfer learning and presents an analysis of effects of gender in transfer.

Download fulltext PDF



Author: Ekin Gedik - Hayley Hung

Source: https://link.springer.com/



DOWNLOAD PDF




Related documents