Pose Transformers (POTR): Human Motion Prediction with Non-Autoregressive Transformers

TL;DR

This paper introduces a non-autoregressive Transformer model for human motion prediction that predicts pose sequences in parallel, reducing computation and error accumulation, and demonstrates competitive results especially for short-term predictions.

Contribution

The paper presents a novel non-autoregressive Transformer framework for human motion prediction, including sequence decoding, activity classification, and competitive performance evaluation.

Findings

Achieves competitive results on public datasets.

Performs better on short-term predictions.

Reduces computational complexity compared to RNN-based methods.

Abstract

We propose to leverage Transformer architectures for non-autoregressive human motion prediction. Our approach decodes elements in parallel from a query sequence, instead of conditioning on previous predictions such as instate-of-the-art RNN-based approaches. In such a way our approach is less computational intensive and potentially avoids error accumulation to long term elements in the sequence. In that context, our contributions are fourfold: (i) we frame human motion prediction as a sequence-to-sequence problem and propose a non-autoregressive Transformer to infer the sequences of poses in parallel; (ii) we propose to decode sequences of 3D poses from a query sequence generated in advance with elements from the input sequence;(iii) we propose to perform skeleton-based activity classification from the encoder memory, in the hope that identifying the activity can improve predictions;(iv) we show that despite its simplicity, our approach achieves competitive results in two public datasets, although surprisingly more for short term predictions rather than for long term ones.