Continuous Human Action Recognition for Human-Machine Interaction: A Review

With advances in data-driven machine learning research, a wide variety of prediction models have been proposed to capture spatio-temporal features for the analysis video streams. Recognising actions and detecting action transitions within an input are challenging but necessary tasks applications that require real-time human-machine interaction. By reviewing large body recent related work literature, we thoroughly analyse, explain, compare segmentation methods provide details on feature extraction strategies used most state-of-the-art methods. We cover impact performance object detection tracking techniques human methodologies. investigate application such real-world scenarios discuss several limitations key research directions towards improving interpretability, generalisation, optimisation, deployment.