EG-HRNet: An Efficient High-Resolution Network Using Ghost-Modules for Human Pose Estimation

Abstract

As multi-scale features are necessary for human pose estimation tasks, high-resolution networks are widely applied. To improve efficiency, lightweight modules are proposed to replace costly point-wise convolutions in high-resolution networks, including channel weighting and spatial weighting methods. However, they fail to maintain the consistency of weights and capture global spatial information. To address these problems, we present a Grouped lightweight High-Resolution Network (Greit-HRNet), in which we propose a Greit block including a group method Grouped Channel Weighting (GCW) and a spatial weighting method Global Spatial Weighting (GSW). GCW modules group conditional channel weighting to make weights stable and maintain the highresolution features with the deepening of the network, while GSW modules effectively extract global spatial information and exchange information across channels. In addition, we apply the Large Kernel Attention (LKA) method to improve the whole efficiency of our Greit-HRNet. Our experiments on both MS-COCO and MPII human pose estimation datasets demonstrate the superior performance of our Greit-HRNet, outperforming other state-of-the-art lightweight networks.

Research on conflict detection model for taxi‐in process on the apron based on aircraft wingtip keypoint detection

“…ResNet uses the shortcut structure to continuously deepen the network, thereby effectively solving the problem of accuracy decline when the network depth is high and has certain advantages in training. The use of HRNet networks allows a high resolution to be maintained in the aircraft feature extraction process and works by connecting the multi-resolution subnetworks in parallel [31].…”