The performance of multi-person pose estimation has significantly improved with the development of deep convolutional neural networks. However, two challenging issues are still ignored but are key factors causing deterioration in the keypoint localization. These two issues are scale variation of human body parts and huge information loss caused by consecutive striding in multiple upsampling. In this paper, we present a novel network named 'Scale-aware attention-based multi-resolution representation network' (SaMr-Net) which targets to make the proposed method against scale variation and prevent the detail information loss in upsampling, leading more precisely keypoint estimation. The proposed architecture adopts the highresolution network (HRNet) as the backbone, we first introduce dilated convolution into the backbone to expand the receptive field. Then, attention-based multi-scale feature fusion module is devised to modify the exchange units in the HRNet, allowing the network to learn the weights of each fusion component. Finally, we design a scale-aware keypoint regressor model that gradually integrates features from low to high resolution, enhancing the invariance in different scales of pose parts keypoint estimation. We demonstrate the superiority of the proposed algorithm over two benchmark datasets: (1) the MS COCO keypoint benchmark, and (2) the MPII human pose dataset. The comparison shows that our approach achieves superior results. KeywordsMulti-person pose estimation • Scale-aware attention • Multi-scale feature fusion Communicated by Q. Tian.

No abstract

No abstract

Along with the continuous development of wireless communication technology and popularization of multimode terminal, the seamless fusion of various heterogeneous wireless networks has gradually become the trend of network development. The diversified forms of network access enable the multi-mode terminal to dynamically select the access network with the highest utility value to obtain network service according to specific standard. The study selects a typical heterogeneous network scenario, abstracts users' network selection process to be group game model, adopts the evolutionary methods, and then studies user's network selection process through replicator dynamics. On the basis of it, it puts forwards a kind of network selection algorithm in the view of replicator dynamics and verifies its algorithm performance through simulation results, and then analyzes the impact of different factors on evolutionary equilibrium in detail.

No abstract