Generalized Zero Shot Skeletal Action

Generalized Zero-Shot Skeletal Action Recognition is a sub-task in the field of computer vision that aims to recognize unseen action categories through training data of known categories. This task utilizes skeletal data as input and achieves effective recognition and classification of new actions through cross-modal mapping and semantic embedding techniques. Its goal is to enhance the model's generalization ability, reduce the reliance on large-scale annotated data, and thus has significant application value in practical scenarios such as human-computer interaction, behavior analysis, and intelligent surveillance.