Pose as a Modality: A Psychology-Inspired Network for Personality Recognition with a New Multimodal Dataset

TL;DR

This work addresses the problem of accurately predicting Big Five personality traits from multimodal data by introducing a frontal full-body pose dataset and a psychology-inspired network, PINet. PINet combines Multimodal Feature Awareness, Multimodal Feature Interaction, and Psychology-Informed Modality Correlation Loss to exploit pose alongside facial, audio, text, and frame modalities, aided by Vision Mamba and Fusion Mamba components. The approach yields state-of-the-art performance across multiple baselines, with pose data providing a meaningful contribution and the PIMC Loss enabling trait-specific modality weighting. The study demonstrates the value of integrating psychological theory into AI models and fills a gap in datasets by offering high-quality full-body pose data for personality prediction, with implications for more accurate and interpretable multimodal systems.

Abstract

In recent years, predicting Big Five personality traits from multimodal data has received significant attention in artificial intelligence (AI). However, existing computational models often fail to achieve satisfactory performance. Psychological research has shown a strong correlation between pose and personality traits, yet previous research has largely ignored pose data in computational models. To address this gap, we develop a novel multimodal dataset that incorporates full-body pose data. The dataset includes video recordings of 287 participants completing a virtual interview with 36 questions, along with self-reported Big Five personality scores as labels. To effectively utilize this multimodal data, we introduce the Psychology-Inspired Network (PINet), which consists of three key modules: Multimodal Feature Awareness (MFA), Multimodal Feature Interaction (MFI), and Psychology-Informed Modality Correlation Loss (PIMC Loss). The MFA module leverages the Vision Mamba Block to capture comprehensive visual features related to personality, while the MFI module efficiently fuses the multimodal features. The PIMC Loss, grounded in psychological theory, guides the model to emphasize different modalities for different personality dimensions. Experimental results show that the PINet outperforms several state-of-the-art baseline models. Furthermore, the three modules of PINet contribute almost equally to the model's overall performance. Incorporating pose data significantly enhances the model's performance, with the pose modality ranking mid-level in importance among the five modalities. These findings address the existing gap in personality-related datasets that lack full-body pose data and provide a new approach for improving the accuracy of personality prediction models, highlighting the importance of integrating psychological insights into AI frameworks.

Figures (6)

• Figure 1: Dataset illustration. The first row shows the visualization results using AlphaPose alphapose, and the second row shows the dataset's pose data.
• Figure 2: The illustration of the data collection scenarios.
• Figure 3: (a) The architecture of our PINet. (b) The Vision Mamba Block in MFA. (c) The Fusion Mamba Block in MFI with the symmetrical parts of the Fusion Mamba Block omitted.
• Figure 4: Ablation study with PIMC Loss, MFI and VMB.
• Figure 5: Ablation study with pose data.