Hand Pose Estimation Based on Deep Learning

The problem of 3D hand pose estimation has aroused a lot of attention in computer vision community for long time. It has been studied in computer vision for decades, as it plays a significant role in human-computer interaction such as virtual/augmented reality applications, computer graphics and robotics. Because of the practical value associated with this topic, it regained huge research interests recently due to the emergence of commodity depth cameras. But despite the recent progress in this field, robust and accurate hand pose estimation remains a challenging task due to the large pose variations, the high dimension of hand motion, the highly articulated structure, significant self-occlusion, viewpoint changes and data noises. Besides, real time performance is often desired in many applications. In this work we have tried to make a comparative study of different methods of hand pose estimation introduced recently, we worked on the implementation of our method based on Deep Learning to solve this problem. © 2021, The Author(s), under exclusive license to Springer Nature Switzerland AG.