DoubleHigherNet : coarse-to-fine precise heatmap bottom-up dynamic pose computer intelligent estimation

Tools

Peng, Yiheng and Jiang, Zhichun (2021) DoubleHigherNet : coarse-to-fine precise heatmap bottom-up dynamic pose computer intelligent estimation. Journal of Physics: Conference Series, 2033 (1). 012068. ISSN 1742-6588 (https://doi.org/10.1088/1742-6596/2033/1/012068)

[thumbnail of Peng-Jiang-JPCS-2021-DoubleHigherNet-corase-to-fine-precise-heatmap-bottom-up]
Preview
 Text. Filename: Peng_Jiang_JPCS_2021_DoubleHigherNet_corase_to_fine_precise_heatmap_bottom_up.pdf
Final Published Version
License: Creative Commons Attribution 4.0 logo
Download (707kB)| Preview

Abstract

Accurate keypoint positioning is necessary for bottom-up multi-person pose estimation methods to handle scale variation and crowdedness. In this paper, we present DoubleHigherNet: a novel network learning scale-aware and precise heatmap representation for bottom-up process using double high-resolution feature pyramids and coarse-to-fine training. The two feature pyramids in DoubleHigherNet consists of 1/4 resolution feature and higher-resolution (1/2) maps generated by attention fusion blocks and transposed convolutions. Benefited by the training strategy, muti-resoltion and coarse-fine heatmap aggregation, the proposed approach is able to predict keypoints more accurately so as to perform better on difficult crowded scenes. DoubleHigherNetw32 achieves competitive result on CrowdPose-test, surpassing all the top-down methods and bottom-up SOTA HigherHRNet-w32 (which possesses similar number of params with DoubleHigherNet-w32).

  • Item type:Article
    ID code:78974
    Dates:
    Date
    Event
    5 October 2021
    Published
    13 June 2021
    Published Online
    30 April 2021
    Accepted
    Subjects:Science > Physics
    Department:UNSPECIFIED
    Depositing user: Pure Administrator
    Date deposited:20 Dec 2021 10:57
    Last modified:11 Nov 2024 13:20
    Related URLs:
    URI:https://strathprints.strath.ac.uk/id/eprint/78974
CORE (COnnecting REpositories)