Deep Learning for Head Pose Estimation: A Survey

Abstract Head pose estimation (HPE) is an active and popular area of research. Over the years, many approaches have constantly been developed, leading to a progressive improvement in accuracy; nevertheless, head remains open research topic, especially unconstrained environments. In this paper, we will review increasing amount available datasets modern methodologies used estimate orientation, with special attention deep learning techniques. We discuss evolution field by proposing classification methods, explaining their advantages disadvantages, highlighting different ways techniques context HPE. An in-depth performance comparison discussion presented at end work. also highlight most promising directions for future investigations on topic.