No open-access location is known for this record yet.
No open-access location is known for this record yet.
Understanding the inner psychological world of children through their drawings has long been a cornerstone of developmental and clinical psychology. Traditional assessment techniques, such as the House-Tree-Person (HTP) test, rely heavily on expert interpretation, which is often subjective, time-consuming, and inconsistent across evaluators. In this research, we propose a novel AI-powered framework that fuses deep learning and natural language processing to automate the psychological interpretation of children’s HTP drawings. The system employs ResNet-34, a convolutional neural network (CNN), for advanced image analysis to detect and interpret nuanced visual patterns in the drawings. These patterns are then transformed into coherent, human-like psychological insights using a transformer-based language generation model, specifically BART. Unlike static, template-driven systems, our approach dynamically generates rich, context-aware psychological narratives that align with clinical reasoning. This interdisciplinary methodology presents a significant step forward in augmenting traditional psychological evaluation with scalable, interpretable, and data-driven intelligence.
Loading…
Imported from OpenAlex (CC0) — nothing here is generated.