English

Graph-Based Multimodal and Multi-view Alignment for Keystep Recognition

Computer Vision and Pattern Recognition 2026-02-11 v2

Abstract

Egocentric videos capture scenes from a wearer's viewpoint, resulting in dynamic backgrounds, frequent motion, and occlusions, posing challenges to accurate keystep recognition. We propose a flexible graph-learning framework for fine-grained keystep recognition that is able to effectively leverage long-term dependencies in egocentric videos, and leverage alignment between egocentric and exocentric videos during training for improved inference on egocentric videos. Our approach consists of constructing a graph where each video clip of the egocentric video corresponds to a node. During training, we consider each clip of each exocentric video (if available) as additional nodes. We examine several strategies to define connections across these nodes and pose keystep recognition as a node classification task on the constructed graphs. We perform extensive experiments on the Ego-Exo4D dataset and show that our proposed flexible graph-based framework notably outperforms existing methods by more than 12 points in accuracy. Furthermore, the constructed graphs are sparse and compute efficient. We also present a study examining on harnessing several multimodal features, including narrations, depth, and object class labels, on a heterogeneous graph and discuss their corresponding contribution to the keystep recognition performance.

Keywords

Cite

@article{arxiv.2501.04121,
  title  = {Graph-Based Multimodal and Multi-view Alignment for Keystep Recognition},
  author = {Julia Lee Romero and Kyle Min and Subarna Tripathi and Morteza Karimzadeh},
  journal= {arXiv preprint arXiv:2501.04121},
  year   = {2026}
}

Comments

We expanded the paper and resubmitted as a separate submission to arXiv. This submission is outdated and readers can refer to arXiv:2506.01102

R2 v1 2026-06-28T20:59:15.259Z