TY - GEN
T1 - Resolution irrelevant encoding and difficulty balanced loss based network independent supervision for multi-person pose estimation
AU - Liu, Haiyang
AU - Luo, Dingli
AU - Du, Songlin
AU - Ikenaga, Takeshi
N1 - Funding Information:
This work was supported by Waseda University Grant for Special Research Projects (2019C-581).
Funding Information:
[13] G. Huang, Z. Liu, L. Van Der Maaten, and K. Q. Weinberger, “Densely connected convolutional networks,” in Proceedings of the IEEE confer-ence on computer vision and pattern recognition, 2017, pp. 4700–4708. This work was supported by Waseda University Grant for Z. Cao, T. Simon, S.-E. Wei, and Y. Sheikh, “Realtime multi-person Special Research Projects (2019C-581). IEEEConference onComputerVision andPatternRecognition,2017,2d poseestimationusingpart affinityfields,”inProceedingsofthe
Publisher Copyright:
© 2020 IEEE.
PY - 2020/6
Y1 - 2020/6
N2 - Sustainable efforts are made to improve the accuracy performance in multi-person pose estimation, but the current accuracy is still not enough for real-world applications. Besides, most improvement approaches are designed for special basement networks and ignore the speed performance, which results in limited applicability and low cost-performance. This paper proposes two network independent supervision: Resolution Irrelevant Encoding and Difficulty Balanced Loss. The proposed methods reorganize task representatives, the loss calculation method, and the loss punishment ratio in one-stage pose estimation frameworks to improve the joints' location accuracy with general applicability and high computational efficiency. Resolution Irrelevant Encoding fuses heatmaps and proposed inner block offsets to fix pixel-level joints positions without resolution limitations. To improve network training efficiency, Difficulty Balanced Loss adjusts loss weight in spatial and sequential aspects. On the MS COCO keypoints detection benchmark, the mAP of OpenPose trained with our proposals outperforms the OpenPose baseline over 4.9%.
AB - Sustainable efforts are made to improve the accuracy performance in multi-person pose estimation, but the current accuracy is still not enough for real-world applications. Besides, most improvement approaches are designed for special basement networks and ignore the speed performance, which results in limited applicability and low cost-performance. This paper proposes two network independent supervision: Resolution Irrelevant Encoding and Difficulty Balanced Loss. The proposed methods reorganize task representatives, the loss calculation method, and the loss punishment ratio in one-stage pose estimation frameworks to improve the joints' location accuracy with general applicability and high computational efficiency. Resolution Irrelevant Encoding fuses heatmaps and proposed inner block offsets to fix pixel-level joints positions without resolution limitations. To improve network training efficiency, Difficulty Balanced Loss adjusts loss weight in spatial and sequential aspects. On the MS COCO keypoints detection benchmark, the mAP of OpenPose trained with our proposals outperforms the OpenPose baseline over 4.9%.
KW - Human pose estimation
KW - Network independent
KW - Supervision strategy
UR - http://www.scopus.com/inward/record.url?scp=85091396118&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85091396118&partnerID=8YFLogxK
U2 - 10.1109/HSI49210.2020.9142625
DO - 10.1109/HSI49210.2020.9142625
M3 - Conference contribution
AN - SCOPUS:85091396118
T3 - International Conference on Human System Interaction, HSI
SP - 112
EP - 117
BT - Proceedings - 2020 13th International Conference on Human System Interaction, HSI 2020
PB - IEEE Computer Society
T2 - 13th International Conference on Human System Interaction, HSI 2020
Y2 - 6 June 2020 through 8 June 2020
ER -