Towards Arbitrary Motion Completing via Hierarchical Continuous Representation

TL;DR

This paper introduces a hierarchical implicit neural representation framework for continuous human motion modeling, enabling interpolation, inbetweening, and extrapolation at arbitrary frame rates with high accuracy.

Contribution

It proposes a novel parametric activation-induced hierarchical INR framework that captures complex motion patterns across multiple temporal scales.

Findings

Effective motion interpolation and extrapolation demonstrated

High accuracy in representing intricate motion behaviors

Robust performance across benchmark datasets

Abstract

Physical motions are inherently continuous, and higher camera frame rates typically contribute to improved smoothness and temporal coherence. For the first time, we explore continuous representations of human motion sequences, featuring the ability to interpolate, inbetween, and even extrapolate any input motion sequences at arbitrary frame rates. To achieve this, we propose a novel parametric activation-induced hierarchical implicit representation framework, referred to as NAME, based on Implicit Neural Representations (INRs). Our method introduces a hierarchical temporal encoding mechanism that extracts features from motion sequences at multiple temporal scales, enabling effective capture of intricate temporal patterns. Additionally, we integrate a custom parametric activation function, powered by Fourier transformations, into the MLP-based decoder to enhance the expressiveness of the continuous representation. This parametric formulation significantly augments the model's ability to represent complex motion behaviors with high accuracy. Extensive evaluations across several benchmark datasets demonstrate the effectiveness and robustness of our proposed approach.