.png){kind=icon}
Return
An efficient sparse pruning method for human pose estimation
.png){kind=icon}
OA
AI
DOI：10.1080/09540091.2021.2012423.png)
.png){kind=icon}
.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Human pose estimation (HPE) is crucial for computer vision (CV). Moreover, it's a vital step for computers to understand human actions and behaviours. However, the huge number of parameters and calculations in the HPE model have brought big challenges to deploy to resource-constrained mobile devices. Aiming to overcome the challenge, we propose a sparse pruning method (SPM) for the HPE model. First, L1 regularisation is added in the training phase of the original model, and network parameters of the convolution layers (CLs) and batch normalisation layers (BNLs) are sparsely trained to obtain a network structure with sparse weights. We then combine the sparse weights of filters with the scaling parameters of the BNLs to determine their importance. Finally, the structured pruning method is used to prune the sparse filters and corresponding channels. SPM can reduce the number of model parameters and calculations without affecting precision. Promising results indicate that SPM outperforms other advanced pruning methods.
Keywords:
Human pose estimation
computer vision
sparse pruning method
scaling parameters
structured pruning method
Journal
IF:
3.4
Papers:
855
Citations:
1.5K
Organization
View more organizations
Cited Papers
ShareSaveSemantic segmentation of high-resolution remote sensing images using fully convolutional network with adaptive threshold.png)
CONNECTION SCIENCE
CONNECTION SCIENCEIF3.4
PREAI
ShareSaveShareSave
ShareSaveApplication of Kalman Filter Method Based on Water Level Factors in the Dam Deformation Forecast
PREAI
ShareSave View more