HEMlets Pose: Learning Part-Centric Heatmap Triplets for Accurate 3D Human Pose Estimation
About
Estimating 3D human pose from a single image is a challenging task. This work attempts to address the uncertainty of lifting the detected 2D joints to the 3D space by introducing an intermediate state - Part-Centric Heatmap Triplets (HEMlets), which shortens the gap between the 2D observation and the 3D interpretation. The HEMlets utilize three joint-heatmaps to represent the relative depth information of the end-joints for each skeletal body part. In our approach, a Convolutional Network (ConvNet) is first trained to predict HEMlests from the input image, followed by a volumetric joint-heatmap regression. We leverage on the integral operation to extract the joint locations from the volumetric heatmaps, guaranteeing end-to-end learning. Despite the simplicity of the network design, the quantitative comparisons show a significant performance improvement over the best-of-grade method (by 20% on Human3.6M). The proposed method naturally supports training with "in-the-wild" images, where only weakly-annotated relative depth information of skeletal joints is available. This further improves the generalization ability of our model, as validated by qualitative comparisons on outdoor images.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| 3D Human Pose Estimation | MPI-INF-3DHP (test) | PCK75.3 | 559 | |
| 3D Human Pose Estimation | Human3.6M (Protocol #1) | MPJPE (Avg.)39.9 | 440 | |
| 3D Human Pose Estimation | Human3.6M Protocol 1 (test) | Dir. Error (Protocol 1)34.4 | 183 | |
| 3D Human Pose Estimation | Human3.6M (subjects 9 and 11) | -- | 180 | |
| 3D Human Pose Estimation | Human3.6M v1 (test) | Avg Performance39.9 | 58 | |
| 3D Human Pose Estimation | HumanEva | Walk S1 Error13.5 | 32 | |
| 3D Human Pose Estimation | Human3.6M Protocol 2 (subjects 9 and 11) | Avg Error (mm)32.1 | 19 | |
| 3D Human Pose Estimation | MPI-INF-3DHP Universal, height-normalized skeletons 1.0/2.0 (test) | -- | 8 |