Skip to main navigation Skip to search Skip to main content

Somatisation Disorder Recognition by Stream Fusion with WavLM and Enhanced ResNet

  • Zhijing Cao
  • , Lin Shen
  • , Liuxian Ma
  • , Xiaoxi Liu
  • , Haojie Zhang
  • , Yongxin Zhang
  • , Yilu Deng
  • , Kun Qian*
  • , Ruolan Huang*
  • , Toru Nakamura
  • , Bin Hu*
  • , Björn W. Schuller
  • , Yoshiharu Yamamoto
  • *Corresponding author for this work
  • Key Laboratory of Brain Health Intelligent Evaluation and Intervention (BIT)
  • Beijing Institute of Technology
  • The University of Osaka
  • North China Electric Power University
  • Nanfang Hospital
  • Shenzhen University
  • Imperial College London
  • Technical University of Munich
  • The University of Tokyo

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Somatisation Disorder (SD) is a serious mental health condition with the difficulty of clinical diagnosis due to the limitations of subjective assessment methods and the lack of objective biomarkers. In this work, we introduce an effective multi-stream fusion strategy and an improved ResNet model for SD recognition. The model combines the WavLM and ResNet architectures. The results show that our model achieves 36.8 % accuracy on the considered quaternary classification task. Based on the proposed model, we likewise realise effective and robust recognition of SD.

Original languageEnglish
Title of host publicationGCCE 2025 - 2025 IEEE 14th Global Conference on Consumer Electronics
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages725-726
Number of pages2
ISBN (Electronic)9798331524166
DOIs
Publication statusPublished - 2025
Externally publishedYes
Event14th IEEE Global Conference on Consumer Electronics, GCCE 2025 - Osaka, Japan
Duration: 23 Sept 202526 Sept 2025

Publication series

NameGCCE 2025 - 2025 IEEE 14th Global Conference on Consumer Electronics

Conference

Conference14th IEEE Global Conference on Consumer Electronics, GCCE 2025
Country/TerritoryJapan
CityOsaka
Period23/09/2526/09/25

Keywords

  • Audio
  • Digital Mental Health
  • Multi-stream Fusion
  • Somatisation Disorder
  • Speech

Fingerprint

Dive into the research topics of 'Somatisation Disorder Recognition by Stream Fusion with WavLM and Enhanced ResNet'. Together they form a unique fingerprint.

Cite this