Yongqiong Zhu
ABSTRACT
Aiming at the noise caused by missing marker data in optical human motion capture, an improved LSTNet neural network model was proposed in this paper, which decomposed the noise prediction into linear part and nonlinear part. In the nonlinear part, convolutional neural network and recurrent neural network are used to deal with periodic prediction, and LSTM is used to replace the gated recurrent unit GRU to enhance memory function. The linear part uses autoregressive models to deal with aperiodic predictions. Finally, the loss function based on the position of markers is constructed to improve the prediction accuracy. The simulation results show that the proposed denoising technique can obtain lower reconstruction error and strong robustness, and the reconstructed motion sequence is very close to the real motion sequence.
- Gløersen, Ø., & Federolf, P. Predicting missing marker trajectories in human motion data using marker intercorrelations. PLoS ONE, 2016:11(3).Google Scholar
- Perepichka, M., Holden, D., Mudur, S. P., & Popa, T. (2019). Robust marker trajectory repair for MOCAP using kinematic reference. Proceedings - MIG 2019: ACM Conference on Motion, Interaction, and Games.Google ScholarDigital Library
- Xia, G., Sun, H., Zhang, G., & Feng, L. Human motion recovery jointly utilizing statistical and kinematic information. Information Sciences, 2016: 339, 189–205.Google Scholar
- Burke, M., & Lasenby, J. Estimating missing marker positions using low dimensional Kalman smoothing. Journal of Biomechanics, 2016:49(9), 1854–1858.Google Scholar
- Baumann, Jan, Björn Krüger, Arno Zinke, and Andreas Weber. Data-Driven Completion of Motion Capture Data. Vriphys, pp. 111-118. 2011.Google Scholar
- Judith Bütepage, Michael Black, Danica Kragic, and Hedvig Kjellström. Deep representation learning for human motion prediction and classification. In IEEE Conference on Computer Vision and Pattern Recognition, 2017.Google Scholar
- Mall, U., Roshan Lal, G., Chaudhuri, S., & Chaudhuri, P. A deep recurrent framework for cleaning motion capture data. ArXiv, Figure 1,2017.Google Scholar
- Kucherenko, T., Beskow, J., & Kjellström, H. A Neural Network Approach to Missing Marker Reconstruction in Human Motion Capture. 2018.Google Scholar
- Holden, Daniel. Robust solving of optical motion capture data by denoising. ACM Transactions on Graphics (TOG) 37.4 (2018): 1-12.Google Scholar
- Li, Shujie, "Bidirectional recurrent autoencoder for 3D skeleton motion data refinement." Computers & Graphics 81 (2019): 92-103.Google ScholarDigital Library
- Cui, Qiongjie, Huaijiang Sun, Yupeng Li, and Yue Kong. "A Deep Bi-directional Attention Network for Human Motion Recovery." In IJCAI, pp. 701-707. 2019.Google Scholar
- Lai, Guokun, "Modeling long-and short-term temporal patterns with deep neural networks." The 41st International ACM SIGIR Conference on Research & Development in Information Retrieval. 2018.Google Scholar
Index Terms
- Predicting missing markers in mocap data using LSTNet
Recommendations
Predicting and Analyzing Educational Data Using Neural Networks
ICIIP '23: Proceedings of the 2023 8th International Conference on Intelligent Information ProcessingThis research aims to explore the use of neural networks to predict educational data and analyze the accuracy of predictions and influencing factors. By collecting specific student data, applying neural network models for prediction, and generating ...
Predicting missing markers in human motion capture using l1-sparse representation
Missing marker problem is very common in human motion capture. In contrast to most current methods which handle this problem based on trying to learn a reliable predictor from the observations, we consider it from the perspective of sparse ...
Predicting time series using neural networks with wavelet-based denoising layers
To avoid the need to pre-process noisy data, two special denoising layers based on wavelet multiresolution analysis have been integrated into layered neural networks. A gradient-based learning algorithm has been developed that uses the same cost ...
Comments