Action recognition (movement science) is a field of study within computer vision and human motion analysis that focuses on the computational processes and algorithms used to automatically identify, classify, and understand sequences of human movements or behaviors, referred to as actions, typically from visual data or other forms of motion capture. Drawing upon techniques from areas such as deep learning, it investigates methods for extracting relevant features from motion data and mapping them to predefined action categories. This concept is central to applications involving automated analysis of human activity, performance evaluation, human-computer interaction, and behavioral understanding within the context of movement science.