EgoLM: Multi-Modal Language Model of Egocentric Motions

TL;DR

EgoLM is a multi-modal language model that leverages egocentric videos and sensor data to improve motion tracking and understanding by integrating natural language processing with motion analysis.

Contribution

It introduces a novel framework that models the joint distribution of egocentric motions and language using large language models, enabling versatile multi-modal egocentric learning.

Findings

EgoLM outperforms existing methods in egocentric motion understanding.

The framework effectively integrates multi-modal inputs for improved disambiguation.

Experiments demonstrate EgoLM's versatility across large-scale datasets.

Abstract

As the prevalence of wearable devices, learning egocentric motions becomes essential to develop contextual AI. In this work, we present EgoLM, a versatile framework that tracks and understands egocentric motions from multi-modal inputs, e.g., egocentric videos and motion sensors. EgoLM exploits rich contexts for the disambiguation of egomotion tracking and understanding, which are ill-posed under single modality conditions. To facilitate the versatile and multi-modal framework, our key insight is to model the joint distribution of egocentric motions and natural languages using large language models (LLM). Multi-modal sensor inputs are encoded and projected to the joint latent space of language models, and used to prompt motion generation or text generation for egomotion tracking or understanding, respectively. Extensive experiments on large-scale multi-modal human motion dataset validate the effectiveness of EgoLM as a generalist model for universal egocentric learning.