Distribution-Aware Coordinate Representation for Human Pose Estimation

By Feng Zhang, Xiatian Zhu, Hanbin Dai, Mao Ye, Ce Zhu. Computer Vision and Pattern Recognition, 2019

527 citations recorded by the source corpus.

While being the de facto standard coordinate representation for human pose estimation, heatmap has not been investigated in-depth. This work fills this gap. For the first time, we find that the process of decoding the predicted heatmaps into the final joint coordinates in the original image space is surprisingly significant for the performance. We further probe the design limitations of the standard coordinate decoding method, and propose a more principled distributionaware decoding method. Also, we improve the standard coordinate encoding process (i.e. transforming ground-truth coordinates to heatmaps) by generating unbiased/accurate heatmaps. Taking the two together, we formulate a novel Distribution-Aware coordinate Representation of Keypoints (DARK) method. Serving as a model-agnostic plug-in, DARK brings about significant performance boost to existing human pose estimation models. Extensive experiments show that DARK yields the best results on two common benchmarks, MPII and COCO. Besides, DARK achieves the 2nd place entry in the ICCV 2019 COCO Keypoints Challenge. The code is available online.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/7cc3fccd3f3d3e137f6cd80b75db2c640cf6f8d8

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

Citations in the corpus

TokenPose: Learning Keypoint Tokens for Human Pose Estimation (2021, methodology): “To alleviate the quantisation error, the welldesigned coordinate decoding strategy [42] is adopted.”
Polarized Self-Attention: Towards High-quality Pixel-wise Regression (2021, methodology): “6M Dark-Pose [56] HRNet-W48 384× 288 76.”
HRFormer: High-Resolution Transformer for Dense Prediction (2021, methodology): “We believe that our HRFormer-B could achieve better results by exploiting either UDP or DARK scheme.”
TransPose: Keypoint Localization via Transformer (2020, methodology): “We also adopt the coordinate decoding strategy proposed by [65] to reduce the quantisation error when decoding from downscaled heatmaps.”
Empowering Things With Intelligence: A Survey of the Progress, Challenges, and Opportunities in Artificial Intelligence of Things (2020, unknown): “techniques [107], [114].”
Pose2Mesh: Graph Convolutional Network for 3D Human Pose and Mesh Recovery from a 2D Human Pose (2020, methodology): “In addition, we used 2D human poses estimated from DarkPose [65] as input for 3DPW evaluation, which improved HRNet [56] recently.”
Bottom-Up Human Pose Estimation Via Disentangled Keypoint Regression (2021, background): “Other developments include improving the keypoint localization from the heatmap [22, 73], refining pose estimation [19, 39], better data augmentation [4], developing a multi-task learning architecture combining detection, segmentation and pose estimation [30], and handling the occlusion issue [34, 5”
Pose Recognition with Cascade Transformers (2021, methodology): “DARK [40] presents Taylor-expansion based coordinate decoding and unbiased sub-pixel centered coordinate encoding.”
RTMPose: Real-Time Multi-Person Pose Estimation based on MMPose (2023, methodology): “Compared with heatmap-based algorithms [24, 59, 61, 65], the SimCC-based algorithm achieves competitive accuracy with lower computational effort.”
Deep Learning-based Human Pose Estimation: A Survey (2020, unknown): “3 – – 2020 [283] no HRNet-W48 384 × 288 76.”