RECENT ADVANCES IN HUMAN POSE ESTIMATION: DEEP LEARNING APPROACHES AND REAL-TIME APPLICATIONS
Keywords:
Human Pose Estimation, Deep Learning, Real-Time Processing, Multi-Person Pose Estimation, Edge ComputingAbstract
This comprehensive article explores recent advances in human pose estimation (HPE), a critical computer vision task with wide-ranging applications. The article traces the evolution from traditional methods to cutting-edge deep learning approaches, highlighting the transformative impact of convolutional neural networks and transformer-based architectures. It examines state-of-the-art models such as RTMPose, HRNet, and DEKR, detailing their innovative features and performance improvements. The review discusses significant progress in multi-person pose estimation, real-time processing, and performance in challenging environments. Applications across diverse fields including sports analytics, healthcare, robotics, and augmented reality are explored, with a focus on edge device implementation and the potential for lightweight, adaptable models. The article also addresses ongoing challenges in the field, such as improving accuracy in complex scenarios, enhancing real-time performance, and navigating privacy and ethical concerns. By synthesizing current research and identifying future directions, this review provides a comprehensive overview of the rapidly evolving landscape of human pose estimation, its practical implications, and the potential for further advancements that could revolutionize human-computer interaction and our understanding of human movement.
References
A. Toshev and C. Szegedy, "DeepPose: Human Pose Estimation via Deep Neural Networks," in 2014 IEEE Conference on Computer Vision and Pattern Recognition, Columbus, OH, USA, 2014, pp. 1653-1660, doi: 10.1109/CVPR.2014.214. [Online]. Available: https://ieeexplore.ieee.org/document/6909610
P. F. Felzenszwalb and D. P. Huttenlocher, "Pictorial Structures for Object Recognition," International Journal of Computer Vision, vol. 61, no. 1, pp. 55-79, 2005. [Online]. Available: https://link.springer.com/article/10.1023/B:VISI.0000042934.15159.49
A. Newell, K. Yang, and J. Deng, "Stacked Hourglass Networks for Human Pose Estimation," in Computer Vision – ECCV 2016, Cham, 2016, pp. 483-499. [Online]. Available: https://link.springer.com/chapter/10.1007/978-3-319-46484-8_29
Z. Cao, G. Hidalgo, T. Simon, S. -E. Wei and Y. Sheikh, "OpenPose: Realtime Multi-Person 2D Pose Estimation Using Part Affinity Fields," in IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 43, no. 1, pp. 172-186, 1 Jan. 2021, doi: 10.1109/TPAMI.2019.2929257. [Online]. Available: https://ieeexplore.ieee.org/document/8765346
J. Yao, et al., "RTMPOSE: Real-Time Multi-Person Pose Estimation based on MMPose," arXiv preprint arXiv:2303.07399, 2023. [Online]. Available: https://arxiv.org/abs/2303.07399
K. Sun, B. Xiao, D. Liu and J. Wang, "Deep High-Resolution Representation Learning for Human Pose Estimation," 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Long Beach, CA, USA, 2019, pp. 5686-5696, doi: 10.1109/CVPR.2019.00584. [Online]. Available: https://arxiv.org/abs/1902.09212
W. Zhang, Z. Zhu, H. Zhu, J. Shi and C. C. Loy, "MobilePose: Real-Time Pose Estimation for Unseen Subjects with Weak Supervision," 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), New Orleans, LA, USA, 2022, pp. 13523-13532, doi: 10.1109/CVPR52688.2022.01315. [Online]. Available: https://arxiv.org/abs/2003.03522
D. C. Luvizon, D. Picard and H. Tabia, "Multi-task Deep Learning for Real-Time 3D Human Pose Estimation and Action Recognition," in IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 43, no. 8, pp. 2752-2764, 1 Aug. 2021, doi: 10.1109/TPAMI.2020.2978534. [Online]. Available: https://arxiv.org/abs/1912.08077
Luyu Qiu, Jianing Li, Lei Wen, Chi Su, Fei Hao, Chen Jason Zhang, Lei Chen, " XPose: eXplainable Human Pose Estimation" [Online]. Available: https://arxiv.org/abs/2403.12370
