CheoFaMo: A dataset of segmented image sequences for in-depth mood analysis in Vietnamese traditional Chèo

The CheoFaMo dataset is a segmented image sequence resource designed for computational mood recognition in Vietnamese traditional Chèo, consisting of 5,844 labeled segments extracted from seven plays. The dataset captures challenging conditions such as theatrical makeup, dynamic lighting, and motion, making it a suitable benchmark for deep learning models in facial expression and mood analysis. Its integration with the CheoGoogle Ontology further enables structured applications in emotion recognition, psychological study, and educational tools for digitizing the art of Chèo