Back to MLP: A simple baseline for human motion prediction

Conference Article


IEEE Winter Conference on Applications of Computer Vision (WACV)





Doc link


Download the digital copy of the doc pdf document


This paper tackles the problem of human motion prediction, consisting in forecasting future body poses from historically observed sequences. State-of-the-art approaches provide good results, however, they rely on deep learning architectures of arbitrary complexity, such as Recurrent Neural Networks (RNN), Transformers or Graph Convolutional Networks (GCN), typically requiring multiple training stages and more than 2 million parameters. In this paper, we show that, after combining with a series of standard practices, such as applying Discrete Cosine Transform (DCT), predicting residual displacement of joints and optimizing velocity as an auxiliary loss, a light-weight network based on multi-layer perceptrons (MLPs) with only 0.14 million parameters can surpass the state-of-the-art performance. An exhaustive evaluation on the Human3.6M, AMASS, and 3DPW datasets shows that our method, named siMLpe, consistently outperforms all other approaches. We hope that our simple method could serve as a strong baseline for the community and allow re-thinking of the human motion prediction problem. The code is publicly available at


computer vision.

Author keywords

Training, Computer vision, Recurrent neural networks, Deep architecture, Transformers, Discrete cosine transforms, Convolutional neural networks

Scientific reference

W. Guo, Y. Du, X. Shen, V. Lepetit, X. Alameda and F. Moreno-Noguer. Back to MLP: A simple baseline for human motion prediction, 2023 IEEE Winter Conference on Applications of Computer Vision, 2023, Waikoloa, Hawaii, pp. 4798-4808.