Savinay Nagendra | Penn State University (original) (raw)
Uploads
Papers by Savinay Nagendra
Human pose stability analysis is the key to understanding locomotion and control of body equilibr... more Human pose stability analysis is the key to understanding locomotion and control of body equilibrium, with numerous applications in the fields of kinesiology, medicine, and robotics. In Biomechanics, analysis of Center of Pressure (CoP) is used in studies of human postural control and gait. We propose and validate a novel approach to learn CoP dynamics from pose kinematics of a human body to aid stability analysis. More specifically, we propose an end-to-end deep learning architecture to regress foot pressure heatmaps, and hence the CoP locations, from a human pose derived from video. We have collected and utilized a set of long (5min +) choreographed Taiji (Tai Chi) sequences of multiple subjects with synchronized motion capture, foot pressure, and video data. The derived human pose data and corresponding foot pressure maps are used jointly in training a Convolutional neural network with residual architecture, named PressNET. Cross-subject validation results show promising performance of Press-NET, significantly outperforming the baseline method of K-Nearest Neighbors under reasonable sensor noise ranges. We also show that the CoP locations computed from our regressed foot pressure maps are significantly more accurate than those obtained from the baseline approach meeting the expected outcome of lab-based sensors.
Human pose stability analysis is the key to understanding locomotion and control of body equilibr... more Human pose stability analysis is the key to understanding locomotion and control of body equilibrium, with numerous applications in the fields of kinesiology, medicine, and robotics. In Biomechanics, analysis of Center of Pressure (CoP) is used in studies of human postural control and gait. We propose and validate a novel approach to learn CoP dynamics from pose kinematics of a human body to aid stability analysis. More specifically, we propose an end-to-end deep learning architecture to regress foot pressure heatmaps, and hence the CoP locations, from a human pose derived from video. We have collected and utilized a set of long (5min +) choreographed Taiji (Tai Chi) sequences of multiple subjects with synchronized motion capture, foot pressure, and video data. The derived human pose data and corresponding foot pressure maps are used jointly in training a Convolutional neural network with residual architecture, named PressNET. Cross-subject validation results show promising performance of Press-NET, significantly outperforming the baseline method of K-Nearest Neighbors under reasonable sensor noise ranges. We also show that the CoP locations computed from our regressed foot pressure maps are significantly more accurate than those obtained from the baseline approach meeting the expected outcome of lab-based sensors.