Related Experiment Videos
A hybrid CNN-machine learning model for anxiety screening with mandala coloring patterns: Feature extraction and
Se-Ryun Park1, JongHan Kim2, Jinyoung Han3
1Department of Pharmacology, Kosin University College of Medicine, Busan, Republic of Korea.
Acta Psychologica
|July 24, 2026
Summary
This study used a hybrid model to analyze mandala coloring for anxiety assessment, finding visual features can support classification but aren't a standalone tool.
Area of Science:
- Computational psychology and digital art analysis.
- Machine learning applications in mental health assessment.
Background:
- Traditional anxiety assessments often miss behavioral cues.
- Self-report methods may not capture all anxiety indicators.
Purpose of the Study:
- To develop a hybrid convolutional neural network (CNN) and machine learning (ML) model for classifying visual features in mandala images.
- To explore the potential of quantifiable visual features as markers for anxiety severity.
Main Methods:
- Analysis of 1044 digitized mandala images using features like coloring accuracy, color transitions, color entropy, and symmetry.
- Implementation of a hybrid CNN-ML model with a Random Forest (RF) classifier.
- Stratified 5-fold cross-validation was employed to evaluate model performance.
Main Results:
- The Random Forest classifier achieved an overall accuracy of 84.56% in classifying anxiety levels.
- Model performance was limited under class-imbalanced conditions, with low sensitivity (0.216) for the high-anxiety group.
- Quantifiable visual features showed potential as exploratory markers for anxiety classification.
Conclusions:
- Visual features from mandala coloring can supplement anxiety severity classification.
- The developed computational approach offers a framework for integrating art-based methods with AI for assessment support.
- Further research is needed to refine this method as a diagnostic tool.