Predict the writer's trait emotional intelligence from reproduced calligraphy.

Area of Science:

• Psychology
• Computer Science
• Art Analysis

Background:

• Trait emotional intelligence (EI) relates to emotion regulation.
• Chinese calligraphy emphasizes the connection between character and person.
• The hypothesis is that calligraphy reproductions may reveal a writer's trait EI.

Purpose of the Study:

• To develop and validate a predictive method for trait EI using calligraphy reproductions.
• To investigate the latent information within calligraphy related to personality traits.

Main Methods:

• Construction of a hard pen calligraphy reproduction dataset (48,826 characters, 191 participants).
• Integration of deep learning (Siamese neural network) with aesthetic features.
• Extraction of deep feature differences and combination with handcrafted features for regression analysis.

Main Results:

• The proposed method achieved high prediction accuracy (MAE: 0.463, MSE: 0.462, PCC: 0.730).
• The AI model significantly outperformed human evaluators (MAE: 1.006, MSE: 1.740, PCC: 0.145).
• Calligraphy reproductions contain discernible information about the writer's trait EI.

Conclusions:

• Calligraphy reproductions can be used to predict trait emotional intelligence.
• AI-driven analysis of artistic expression offers novel insights into psychological traits.
• This study bridges computational methods with traditional art forms for personality assessment.