Abstract:Conventional 3D human pose estimation relies on first detecting 2D body keypoints and then solving the 2D to 3D correspondence <a class="link-external link-http" href="http://problem.Despite" rel="external noopener nofollow">this http URL</a> the promising results, this learning paradigm is highly dependent on the quality of the 2D keypoint detector, which is inevitably fragile to occlusions and out-of-image <a class="link-external link-http" href="http://absences.In" rel="external noopener nofollow">this http URL</a> this paper,we propose a novel Pose Orientation Net (PONet) that is able to robustly estimate 3D pose by learning orientations only, hence bypassing the error-prone keypoint detector in the absence of image evidence. For images with partially invisible limbs, PONet estimates the 3D orientation of these limbs by taking advantage of the local image evidence to recover the 3D <a class="link-external link-http" href="http://pose.Moreover" rel="external noopener nofollow">this http URL</a>, PONet is competent to infer full 3D poses even from images with completely invisible limbs, by exploiting the orientation correlation between visible limbs to complement the estimated poses,further improving the robustness of 3D pose <a class="link-external link-http" href="http://estimation.We" rel="external noopener nofollow">this http URL</a> evaluate our method on multiple datasets, including Human3.6M, MPII, MPI-INF-3DHP, and 3DPW. Our method achieves results on par with state-of-the-art techniques in ideal settings, yet significantly eliminates the dependency on keypoint detectors and the corresponding computation burden. In highly challenging scenarios, such as truncation and erasing, our method performs very robustly and yields much superior results as compared to state of the art,demonstrating its potential for real-world applications.

Monocular 3D human pose estimation via sequential second order cone programming

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

Efficient Multi-person Hierarchical 3D Pose Estimation for Autonomous Driving

Marker-Less 3d Human Motion Capture With Monocular Image Sequence And Height-Maps

Motion Imitation of a Humanoid Robot Via Pose Estimation

Skeleton Cluster Tracking for robust multi-view multi-person 3D human pose estimation

Unsupervised Universal Hierarchical Multi-Person 3D Pose Estimation for Natural Scenes

3D Human Pose Estimation using Spatio-Temporal Networks with Explicit Occlusion Training

A Survey on Monocular 3D Human Pose Estimation

Robust Estimation of 3D Human Poses from a Single Image

Sparseness Meets Deepness: 3D Human Pose Estimation from Monocular Video

Robust 3D Human Pose Estimation from Single Images or Video Sequences

PONet: Robust 3D Human Pose Estimation via Learning Orientations Only

Hybrid 3D Human Pose Estimation with Monocular Video and Sparse IMUs

Monocular 3D Human Pose Estimation by Predicting Depth on Joints

Recent Advances in Monocular 2D and 3D Human Pose Estimation: A Deep Learning Perspective

MMDA: Multi-person marginal distribution awareness for monocular 3D pose estimation

MonoCap: Monocular Human Motion Capture Using a CNN Coupled with a Geometric Prior

3D human pose recovery from image by efficient visual feature selection

RSB-Pose: Robust Short-Baseline Binocular 3D Human Pose Estimation with Occlusion Handling

Simultaneous 3-D Human-Motion Tracking and Voxel Reconstruction