Abstract
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.
Paper: [PDF];
Supplementary: [PDF]
Poster: [PDF]
DemoVideo1: [MP4];DemoVideo2: [MP4]
New - Code: [GitHub link]
-----------------------------------------------------------------------------------------------------------------------------------------------------
New - Extended long version
"HEMlets PoSh: Learning Part-Centric Heatmap Triplets for 3D Human Pose and Shape Estimation, " IEEE Trans. on Pattern Analysis and Machine Intelligence (TPAMI), 2021.
Paper: [PDF];
DemoVideo-1: [link];(_3DPW_downtown_crossStreets_00.avi, without using any temporal smoothing)
DemoVideo-2: [link];(_3DPW_downtown_sitOnStairs_00.avi, without using any temporal smoothing)
DemoVideo-3: [link];(_3DPW_downtown_sitOnStairs_00_smplify.avi, without using any temporal smoothing, but using the SMPLify optimization as post-processing.)