Real-time human pose estimation on a smart walker using convolutional neural networks

作者：

Highlights：

• Real-time full-body human pose estimation solution for the ASBGo smart walker.

• Convolutional neural network for 2D keypoint detection with regression to 3D space.

• Information fusion from two rgb+d cameras with non-overlapping, complementary views.

• Exploration of data acquisition, model training, benchmarking and deployment on CPU.

摘要

•Real-time full-body human pose estimation solution for the ASBGo smart walker.•Convolutional neural network for 2D keypoint detection with regression to 3D space.•Information fusion from two rgb+d cameras with non-overlapping, complementary views.•Exploration of data acquisition, model training, benchmarking and deployment on CPU.

论文关键词：rehabilitation,smart walker,computer vision,deep learning,human pose estimation

论文评审过程：Received 14 February 2021, Revised 8 May 2021, Accepted 25 June 2021, Available online 5 July 2021, Version of Record 7 July 2021.

论文官网地址：https://doi.org/10.1016/j.eswa.2021.115498