Skeleton-Based Action Recognition for an Automated Test of Embodied Cognition
摘要
Cognitive assessment is a critical process aimed at evaluating an individual’s cognitive abilities, including memory, attention, problem-solving, and executive functioning. In the context of mental health, cognitive assessment plays a pivotal role in identifying and understanding various neurological disorders such as Alzheimer’s, Schizophrenia, and ADHD. One key aspect of cognitive assessment involves measuring executive functioning, which encompasses higher order cognitive processes responsible for planning, organizing, and regulating behavior. To enhance the efficacy of cognitive assessment in real-world applications, our proposed approach, Mind-In-Action (MIA), focuses on skeleton-based action recognition. MIA integrates a sophisticated pose estimator to extract crucial information from human skeletons, enabling automatic measurement of executive functioning through innovative distance and elbow angle calculations. This methodology introduces three distinct score functions—accuracy score, rhythm score, and functioning score—to comprehensively assess and quantify executive functioning in individuals. Through rigorous evaluations on diverse datasets, our MIA model demonstrates significant advancements over existing methods, showcasing its potential as a valuable tool in the field of cognitive assessment.