Open access
Aug 2026
How Reliable Is Automatic Emotion Classification in Children’s Drawings? A Reproducible Benchmark on a Public Corpus with Calibration and Selective Prediction
A reproducible four-class benchmark (Angry, Fear, Happy, Sad, Sad) is built on a single public corpus of 818 children’s drawings and three transfer-learning regimes under identical stratified five-fold splits with nested model selection are compared.
Hoonhee Lee, Min-woo Kim, Jaewon Kim et al.
· Applied Sciences · 0 citations