“In the field of 3D human pose estimation, researchers have proposed a monocular visual 3D human pose estimation network that integrates multi joint features, effectively improving accuracy and generalization.”