A multi-branch hand pose estimation network with joint-wise feature extraction and fusion

作者：

Highlights：

• Refined feature extraction achieves high accuracy for pose estimation.

• Intermediate supervision in multi-stage feature fusion helps network training.

• Auxiliary points are used as the priori knowledge of hand structure.

摘要

•Refined feature extraction achieves high accuracy for pose estimation.•Intermediate supervision in multi-stage feature fusion helps network training.•Auxiliary points are used as the priori knowledge of hand structure.

论文关键词：Hand pose estimation,Neural network,Human–computer interaction,Depth images

论文评审过程：Received 19 March 2019, Revised 4 October 2019, Accepted 5 November 2019, Available online 8 November 2019, Version of Record 19 November 2019.

论文官网地址：https://doi.org/10.1016/j.image.2019.115692