eduzhai > Applied Sciences > Engineering >

AWR Adaptive Weighting Regression for 3D Hand Pose Estimation

  • king
  • (0) Download
  • 20210506
  • Save

... pages left unread,continue reading

Document pages: 8 pages

Abstract: In this paper, we propose an adaptive weighting regression (AWR) method toleverage the advantages of both detection-based and regression-based methods.Hand joint coordinates are estimated as discrete integration of all pixels indense representation, guided by adaptive weight maps. This learnableaggregation process introduces both dense and joint supervision that allowsend-to-end training and brings adaptability to weight maps, making the networkmore accurate and robust. Comprehensive exploration experiments are conductedto validate the effectiveness and generality of AWR under various experimentalsettings, especially its usefulness for different types of dense representationand input modality. Our method outperforms other state-of-the-art methods onfour publicly available datasets, including NYU, ICVL, MSRA and HANDS 2017dataset.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×