Abstract:In the current digital era, large-scale Artificial Intelligence (AI) models have fundamentally changed the landscape of the AI field. These models excel at extracting complex patterns and information from massive datasets, making them indispensable in a wide range of domains. With the advancement of large-scale AI models, pose estimation has achieved improved performance in managing large data sets and complex computations. However, most existing pose estimation methods suffer from the problem of inaccurate pose graph extraction in complex scenes. Moreover, the existing methods all focus on the message, ignoring the fact that it may not be optimal for the graph structure. This fixed graph structure is difficult to reflect the real dependency relationship between joint points. This study proposes a Character Redetection Model (CRM) to dynamically adjust the foreground extraction strategies to solve these problems. The CRM locates the target position based on the rich spatiotemporal information between the adjacent frames, effectively solving the problem of inaccurate foreground acquisition caused by occlusion and blur in complex scenes. In addition, an Uncertainty Graph Structure (UGS) is designed, which adaptively learns the topological structure of graphs with different Graph Convolutional Network (GCN) layers and skeleton samples in an end-to-end manner and mines the hidden space dependency between data to accurately model the different poses of the characters. The experimental results show that the proposed method achieves a superior performance compared to that of the State-Of-The-Art (SOTA) method, with a percentage of correct keypoints (PCK) value of 94.6% at the Max Planck Institute for Informatics (MPII), along with the mean Average Precision (mAP) values of 83.3% and 82.1% on the PoseTrack2017 and classroom pose datasets, respectively.

Real-Time Pose Estimation Based on ResNet-50 for Rapid Safety Prevention and Accident Detection for Field Workers

Exploring Severe Occlusion: Multi-Person 3D Pose Estimation with Gated Convolution.

OpenPose: Realtime Multi-Person 2D Pose Estimation Using Part Affinity Fields

Research on Real-Time Detection of Safety Harness Wearing of Workshop Personnel Based on YOLOv5 and OpenPose

Live Stream Temporally Embedded 3D Human Body Pose and Shape Estimation

A Real-time Fall Detection System Using ToF Depth Images.

OpenPose : Whole-Body Pose Estimation

ROpenPose: A Rapider OpenPose Model for Astronaut Operation Attitude Detection

Toward Marker-free 3D Pose Estimation in Lifting: A Deep Multi-view Solution

DIR-BHRNet: A Lightweight Network for Real-time Vision-based Multi-person Pose Estimation on Smartphones

Towards Real-Time Head Pose Estimation: Exploring Parameter-Reduced Residual Networks on In-the-wild Datasets

EfficientPose: Scalable single-person pose estimation

Risk Factor Recognition for Automatic Safety Management in Construction Sites Using Fast Deep Convolutional Neural Networks

A Computer Graphics-Based Framework for 3D Pose Estimation of Pedestrians

A Real-Time Predictive Pedestrian Collision Warning Service for Cooperative Intelligent Transportation Systems Using 3D Pose Estimation

Real‐time ergonomic risk assessment in construction using a co‐learning‐powered 3D human pose estimation model

Pose ResNet: 3D Human Pose Estimation Based on Self-Supervision

A Study on Data Collection and Object Detection using Faster R-CNN for Application to Construction Site Safety

Unified End-to-End YOLOv5-HR-TCM Framework for Automatic 2D/3D Human Pose Estimation for Real-Time Applications

Spatiotemporal correlation based self-adaptive pose estimation in complex scenes

Real-time Holistic Robot Pose Estimation with Unknown States