Remembering What Is Important: A Factorised Multi-Head Retrieval and Auxiliary Memory Stabilisation Scheme for Human Motion Prediction
This work addresses the problem of human motion prediction for applications like robotics and animation, offering a novel method that significantly advances the state-of-the-art in this domain-specific area.
The paper tackles the challenge of forecasting future human poses by introducing a deep neural network framework that disentangles subject-specific, task-specific, and auxiliary features from historical motions to query an auxiliary memory, resulting in performance improvements of over 17% on Human3.6M and over 9% on CMU-Mocap datasets compared to state-of-the-art methods.
Humans exhibit complex motions that vary depending on the task that they are performing, the interactions they engage in, as well as subject-specific preferences. Therefore, forecasting future poses based on the history of the previous motions is a challenging task. This paper presents an innovative auxiliary-memory-powered deep neural network framework for the improved modelling of historical knowledge. Specifically, we disentangle subject-specific, task-specific, and other auxiliary information from the observed pose sequences and utilise these factorised features to query the memory. A novel Multi-Head knowledge retrieval scheme leverages these factorised feature embeddings to perform multiple querying operations over the historical observations captured within the auxiliary memory. Moreover, our proposed dynamic masking strategy makes this feature disentanglement process dynamic. Two novel loss functions are introduced to encourage diversity within the auxiliary memory while ensuring the stability of the memory contents, such that it can locate and store salient information that can aid the long-term prediction of future motion, irrespective of data imbalances or the diversity of the input data distribution. With extensive experiments conducted on two public benchmarks, Human3.6M and CMU-Mocap, we demonstrate that these design choices collectively allow the proposed approach to outperform the current state-of-the-art methods by significant margins: $>$ 17\% on the Human3.6M dataset and $>$ 9\% on the CMU-Mocap dataset.