Sparse Representation for 3D Shape Estimation: A Convex Relaxation Approach
arXiv:1509.04309 · doi:10.1109/TPAMI.2016.2605097
Abstract
We investigate the problem of estimating the 3D shape of an object defined by a set of 3D landmarks, given their 2D correspondences in a single image. A successful approach to alleviating the reconstruction ambiguity is the 3D deformable shape model and a sparse representation is often used to capture complex shape variability. But the model inference is still a challenge due to the nonconvexity in optimization resulted from joint estimation of shape and viewpoint. In contrast to prior work that relies on a alternating scheme with solutions depending on initialization, we propose a convex approach to addressing this challenge and develop an efficient algorithm to solve the proposed convex program. Moreover, we propose a robust model to handle gross errors in the 2D correspondences. We demonstrate the exact recovery property of the proposed method, the advantage compared to the nonconvex baseline methods and the applicability to recover 3D human poses and car models from single images.
Extended version of the paper: 3D Shape Estimation from 2D Landmarks: A Convex Relaxation Approach. X. Zhou et al., CVPR, 2015. arXiv admin note: substantial text overlap with arXiv:1411.2942
References in corpus (2)
Cited by in corpus (39)
- VNect: Real-time 3D Human Pose Estimation with a Single RGB Camera
- Lifting from the Deep: Convolutional 3D Pose Estimation from a Single Image
- Graduated Non-Convexity for Robust Spatial Perception: From Non-Minimal Solvers to Global Outlier Rejection
- A simple yet effective baseline for 3d human pose estimation
- SelfPose: 3D Egocentric Pose Estimation from a Headset Mounted Camera
- Unite the People: Closing the Loop Between 3D and 2D Human Representations
- Optimal and Robust Category-level Perception: Object Pose and Shape Estimation from 2D and 3D Semantic Keypoints
- Learning to Estimate 3D Human Pose and Shape from a Single Color Image
- Optimal Pose and Shape Estimation for Category-level 3D Object Perception
- MonoCap: Monocular Human Motion Capture using a CNN Coupled with a Geometric Prior
- HybrIK: A Hybrid Analytical-Neural Inverse Kinematics Solution for 3D Human Pose and Shape Estimation
- 3D Human Pose Estimation from a Single Image via Distance Matrix Regression
- Not All Parts Are Created Equal: 3D Pose Estimation by Modelling Bi-directional Dependencies of Body Parts
- Adversarial Learning of Structure-Aware Fully Convolutional Networks for Landmark Localization
- VoxelPose: Towards Multi-Camera 3D Human Pose Estimation in Wild Environment
- Ego-Pose Estimation and Forecasting as Real-Time PD Control
- Monocular Human Pose and Shape Reconstruction using Part Differentiable Rendering
- To The Point: Correspondence-driven monocular 3D category reconstruction
- Part-level Car Parsing and Reconstruction from Single Street View
- Canonical 3D Deformer Maps: Unifying parametric and non-parametric methods for dense weakly-supervised category reconstruction
- One Ring to Rule Them All: Certifiably Robust Geometric Perception with Outliers
- Ordinal Depth Supervision for 3D Human Pose Estimation
- 3D Human Pose Estimation Based on 2D-3D Consistency with Synchronized Adversarial Training
- Occluded Joints Recovery in 3D Human Pose Estimation based on Distance Matrix
- In Perfect Shape: Certifiably Optimal 3D Shape Reconstruction from 2D Landmarks
- 3D Human Pose Machines with Self-supervised Learning
- Deep Structure for end-to-end inverse rendering
- Keep it SMPL: Automatic Estimation of 3D Human Pose and Shape from a Single Image
- 3D Human Shape Reconstruction from a Polarization Image
- Heuristic Weakly Supervised 3D Human Pose Estimation
- Transformation of Dense and Sparse Text Representations
- Dynamical Pose Estimation
- Monocular 3D Pose Recovery via Nonconvex Sparsity with Theoretical Analysis
- Human Motion Capture Using a Drone
- Deep Pose Consensus Networks
- Kinematic-Structure-Preserved Representation for Unsupervised 3D Human Pose Estimation
- Joint Representation of Multiple Geometric Priors via a Shape Decomposition Model for Single Monocular 3D Pose Estimation
- Birds of a Feather: Capturing Avian Shape Models from Images
- Tree Search Network for Sparse Regression