Abstract
Human motion prediction is an important research frontier, which is a key supporting technology in the fields of human–robot collaboration, automatic driving, etc. As is well known, long-term motion prediction is one most challenging direction. This paper mainly focuses on how to eliminate cumulative errors to overcome the fossilization of long-term motion sequences and aims to improve the reliability of prediction results. This paper proposed an algorithm named “fusion loss learning network,” which is based on gated recurrent unit, to solve the above-mentioned problem. A fusion training method was established by combining the sampling of each step of the GRU unit with true value and output value of each previous step, which helped recover from the errors in the long-term prediction sequences. This method achieved promising results on the Human 3.6 M dataset. The results show that the proposed method could significantly improve the performance of long-term human motion prediction, and the total prediction error is reduced by 7.25% on average.
Similar content being viewed by others
References
Koppula HS, Saxena A (2016) Anticipating human activities using object affordances for reactive robotic response. IEEE Trans Pattern Anal Mach Intell 38(1):14–29
Du X, Vasudevan R, Roberson MJ (2019) Bio-LSTM: a biomechanically inspired recurrent neural network for 3-D pedestrian pose and gait prediction. IEEE RA-L 4(2):1501–1508
Jain A, Zamir AR, Savarese S, Saxena A (2016) Structural-RNN: deep learning on spatio-temporal graphs. Presented at 2016 CVPR. [Online]. Available: https://128.84.21.199/pdf/1511.05298.pdf
Martinez J, Black MJ, Romero J (2017) On human motion prediction using recurrent neural networks. Presented at 2017 CVPR. [Online]. Available: https://arxiv.org/pdf/1705.02445.pdf
Safonova A, Hodgins JK (2009) Synthesizing human motion from intuitive constraints, in studies in computational intelligence. Springer, Berlin, Heidelberg, pp 15–39
Wang JM, Fleet DJ, Hertzmann A (2008) gaussian process dynamical models for human motion. IEEE Trans Pattern Anal Mach Intell 30(2):283–298. https://doi.org/10.1109/TPAMI.2007.1167
Levine S, Wang JM, Haraux A, Popović Z, Koltun V (2012) Continuous character control with low-dimensional embeddings. ACM Trans Graph 31(4):1–10
Pavlovi V, Rehg JM, Maccormick J (2000) Learning switching linear models of human motion. Presented at 2000 NIPS. [Online]. Available: http://dl.acm.org/citation.cfm?id=3008751.3008888
Fragkiadaki K, Levine S, Felsen P, Malik J (2015) Recurrent network models for human dynamics. Presented at 2015 ICCV. [Online]. Available: https://ieeexplore.ieee.org/document/#
Komura T, Habibie I, Holden D, Schwarz J, Yearsley J (2017) A recurrent variational autoencoder for human motion synthesis. Presented at 2017 BMVC. [Online]. Available: https://doi.org/10.5244/c.31.119
Zhou Y, Li Z, Xiao S, He C, Huang Z, Li H (2018) Auto-conditioned recurrent networks for extended complex human motion synthesis. Presented at 2018 ICLR. [Online]. Available: https://arxiv.org/abs/1707.05363
Tang Y, Ma L, Liu W, Zheng WS (2018) Long-term human motion prediction by modeling motion context and enhancing mtion dnamic. Presented at 2018 IJCAI. [Online]. Available: https://www.ijcai.org/proceedings/2018/0130.pdf
Li C, Zhang Z, Lee WS, Lee GH (2018) Convolutional sequence to sequence model for human dynamics. Presented at 2018 CVPR. [Online]. Available: https://ieeexplore.ieee.org/document/#
Guo X, Choi J (2019) Human motion prediction via learning local structure representations and temporal dependencies. Presented at 2019 AAAI. [Online]. Available: https://arxiv.org/pdf/1902.07367v2.pdf
Holden D, Komura T, Saito J (2017) Phase-functioned neural networks for character control. ACM Trans Graph 36(4):1–13
Barsoum E, Kender J, Liu Z (2018) HP-GAN: probabilistic 3D human motion prediction via GAN. Presented at 2018 CVPRW. [Online]. Available: https://ieeexplore.ieee.org/document/#
Gui LY, Wang YX, Liang X, Moura JMF (2018) Adversarial geometry-aware human motion prediction. Presented at 2018 ECCV. [Online]. Available:https://link.springer.com/chapter/10.1007%2F978-3-030-01225-0_48
Graves A (2012) Supervised sequence labelling with recurrent neural networks. Springer, Berlin, Heidelberg
Cho K, Merriënboer BV, Gulcehre C, Bahdanau D, Bougares F, Schwenk H, Bengio Y (2014) Learning phrase representations using RNN encoder-decoder for statistical machine translation. Presented at 2014 EMNLP. [Online]. Available: http://aclweb.org/anthology/D/D14/D14-1179.pdf
Pavllo D, Grangier D, Auli M (2018) QuaterNet: a quaternion-based recurrent model for human motion. Presented at 2018 BMVC. [Online]. Available: http://bmvc2018.org/contents/papers/0675.pdf
Bengio S, Vinyals O, Jaitly N, Shazeer N (2015) Scheduled sampling for sequence prediction with recurrent neural networks. Presented at 2015 NIPS. [Online]. Available: http://arxiv.org/abs/1506.03099
Goyal A, Lamb A, Zhang Y, Zhang S, Courville A, Bengio Y (2016) Professor forcing: a new algorithm for training recurrent networks. Presented at 2016 NIPS. [Online]. Available: http://papers.nips.cc/paper/6098-professor-forcing-a-new-algorithm-for-training-recurrent-networks
Williams RJ (1989) A learning algorithm for continually running fully recurrent neural networks. Neural Comput 1(2):270–280
Ionescu C, Papava D, Olaru V, Sminchisescu C (2014) Human3.6m: large scale datasets and predictive methods for 3d human sensing in natural environments. IEEE Trans Pattern Anal Mach Intell 36(7):1325–1339
Gopalakrishnan A, Mali A, Kifer D, Giles CL, Ororbia AG (2019) A neural temporal model for human motion prediction. Presented at 2019 CVPR. [Online]. Available: http://openaccess.thecvf.com/content_CVPR_2019/papers/Gopalakrishnan_A_Neural_Temporal_Model_for_Human_Motion_Prediction_CVPR_2019_paper.pdf
Acknowledgements
This work was supported in part by the Key Program of NSFC (Grant No. U1908214), Special Project of Central Government Guiding Local Science and Technology Development (Grant No. 2021JH6/10500140), Program for the Liaoning Distinguished Professor, Program for Innovative Research Team in University of Liaoning Province, Dalian and Dalian University, and the Science and Technology Innovation Fund of Dalian (Grant No. 2020JJ25CY001).
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
About this article
Cite this article
Guo, C., Liu, R., Che, C. et al. Fusion learning-based recurrent neural network for human motion prediction. Intel Serv Robotics 15, 245–257 (2022). https://doi.org/10.1007/s11370-021-00403-5
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11370-021-00403-5