ActFusion: a Unified Diffusion Model for Action Segmentation and Anticipation
This work addresses the problem of temporal action analysis in videos for computer vision researchers, offering a unified approach that is incremental by combining existing tasks with a novel training strategy.
The authors tackled the joint problem of action segmentation and anticipation in videos using a unified diffusion model called ActFusion, which achieved state-of-the-art performance on benchmarks like 50 Salads, Breakfast, and GTEA, outperforming task-specific models in both tasks.
Temporal action segmentation and long-term action anticipation are two popular vision tasks for the temporal analysis of actions in videos. Despite apparent relevance and potential complementarity, these two problems have been investigated as separate and distinct tasks. In this work, we tackle these two problems, action segmentation and action anticipation, jointly using a unified diffusion model dubbed ActFusion. The key idea to unification is to train the model to effectively handle both visible and invisible parts of the sequence in an integrated manner; the visible part is for temporal segmentation, and the invisible part is for future anticipation. To this end, we introduce a new anticipative masking strategy during training in which a late part of the video frames is masked as invisible, and learnable tokens replace these frames to learn to predict the invisible future. Experimental results demonstrate the bi-directional benefits between action segmentation and anticipation. ActFusion achieves the state-of-the-art performance across the standard benchmarks of 50 Salads, Breakfast, and GTEA, outperforming task-specific models in both of the two tasks with a single unified model through joint learning.