Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Dynamic Facial Expressions Analysis Based Parkinson's Disease Auxiliary Diagnosis

Xiaochen Huang, Xiaochen Bi, Cuihua Lv, Xin Wang, Haoyan Zhang, Wenjing Jiang, Xin Ma, Yibin Li

TL;DR

This work tackles the challenge of accessible PD auxiliary diagnosis by leveraging dynamic facial expressions (hypomimia). It introduces a CLIP-based multimodal network that fuses visual and textual cues to extract expression intensity from videos, followed by an LSTM classifier, and demonstrates strong cross-validated performance (93.1%). The approach preserves temporal dynamics and is evaluated on the DFEW and PD-FEV datasets, revealing pronounced expression differences between PD patients and healthy controls. While promising for at-home or remote screening, the study notes limitations related to controlled-environment data and model complexity, pointing to future work on diverse data and lightweight deployments.

Abstract

Parkinson's disease (PD), a prevalent neurodegenerative disorder, significantly affects patients' daily functioning and social interactions. To facilitate a more efficient and accessible diagnostic approach for PD, we propose a dynamic facial expression analysis-based PD auxiliary diagnosis method. This method targets hypomimia, a characteristic clinical symptom of PD, by analyzing two manifestations: reduced facial expressivity and facial rigidity, thereby facilitating the diagnosis process. We develop a multimodal facial expression analysis network to extract expression intensity features during patients' performance of various facial expressions. This network leverages the CLIP architecture to integrate visual and textual features while preserving the temporal dynamics of facial expressions. Subsequently, the expression intensity features are processed and input into an LSTM-based classification network for PD diagnosis. Our method achieves an accuracy of 93.1%, outperforming other in-vitro PD diagnostic approaches. This technique offers a more convenient detection method for potential PD patients, improving their diagnostic experience.

Dynamic Facial Expressions Analysis Based Parkinson's Disease Auxiliary Diagnosis

TL;DR

This work tackles the challenge of accessible PD auxiliary diagnosis by leveraging dynamic facial expressions (hypomimia). It introduces a CLIP-based multimodal network that fuses visual and textual cues to extract expression intensity from videos, followed by an LSTM classifier, and demonstrates strong cross-validated performance (93.1%). The approach preserves temporal dynamics and is evaluated on the DFEW and PD-FEV datasets, revealing pronounced expression differences between PD patients and healthy controls. While promising for at-home or remote screening, the study notes limitations related to controlled-environment data and model complexity, pointing to future work on diverse data and lightweight deployments.

Abstract

Parkinson's disease (PD), a prevalent neurodegenerative disorder, significantly affects patients' daily functioning and social interactions. To facilitate a more efficient and accessible diagnostic approach for PD, we propose a dynamic facial expression analysis-based PD auxiliary diagnosis method. This method targets hypomimia, a characteristic clinical symptom of PD, by analyzing two manifestations: reduced facial expressivity and facial rigidity, thereby facilitating the diagnosis process. We develop a multimodal facial expression analysis network to extract expression intensity features during patients' performance of various facial expressions. This network leverages the CLIP architecture to integrate visual and textual features while preserving the temporal dynamics of facial expressions. Subsequently, the expression intensity features are processed and input into an LSTM-based classification network for PD diagnosis. Our method achieves an accuracy of 93.1%, outperforming other in-vitro PD diagnostic approaches. This technique offers a more convenient detection method for potential PD patients, improving their diagnostic experience.

Paper Structure

This paper contains 17 sections, 4 equations, 4 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Methods
  3. Datasets
  4. Dynamic Facial Expression in-the-Wild (DFEW) Dataset
  5. Parkinson’s Disease Facial Expression Videos (PD-FEV) Dataset
  6. Dynamic Facial Expressions Analysis Network
  7. Visual Part
  8. Textual Part
  9. Data Processing and Classification Network
  10. Experiments and Results
  11. Design of Dynamic Facial Expression Analysis Experiments
  12. Ablation Experiment
  13. Experiment Setup
  14. Experiment Results
  15. Conclusion and Disscussion
  16. ...and 2 more sections

Figures (4)

  • Figure 1: Structure of Dynamic Facial Expressions Analysis Based Parkinson's Disease Auxiliary Diagnosis Network.
  • Figure 2: Visualization results of Dynamic Facial Analysis Network in Different Groups.
  • Figure 3: Confusion matrix of dynamic facial expression analysis network.
  • Figure 4: Boxplots Depicting the Varying Intensities of Different Facial Expressions in Each Video.