A path signature approach for speech emotion recognition

Conference item

Automatic speech emotion recognition (SER) remains a difficult task within human-computer interaction, despite increasing interest in the research community. One key challenge is how to effectively integrate short-term characterisation of speech segments with long-term information such as temporal variations. Motivated by the numerical approximation theory of stochastic differential equations (SDEs), we propose the novel use of path signatures. The latter provide a pathwise definition to solve SDEs, for the integration of short speech frames. Furthermore we propose a hierarchical tree structure of path signatures, to capture both global and local information. A simple tree-based convolutional neural network (TBCNN) is used for learning the structural information stemming from dyadic path-tree signatures. Our experimental results on a widely used benchmark dataset demonstrate comparable performance to complex neural network based systems.

Authors: Wang, B; Liakata, M; Ni, H; Lyons, T; Nevado-Holgado, AJ

• Publication status: Published
• Peer review status: Peer reviewed
• Publisher: International Speech Communication Association
• Host title: Proceedings of the Annual Conference of the International Speech Communication Association, Interspeech 2019
• Pages: 1661-1665
• Publication date: 2019-09-15
• Acceptance date: 2019-06-17
• Event title: Interspeech 2019
• Event location: Graz, Austria
• Event website: https://www.interspeech2019.org/
• Event start date: 2019-09-15
• Event end date: 2019-09-19
• DOI: 10.21437/Interspeech.2019-2624
• ISSN: 2308-457X
• Language: English
• Keywords: speech emotion recognition; FFR; convolutional neural network; path signature feature