Future prediction, especially in long-range videos, requires reasoning from current and past observations. In this work, we address questions of temporal extent, scaling, and level of semantic abstraction with a flexible multi-granular temporal aggregation framework. We show that it is possible to achieve state of the art in both next action and dense anticipation with simple techniques such as max-pooling and attention. To demonstrate the anticipation capabilities of our model, we conduct experiments on Breakfast, 50Salads, and EPIC-Kitchens datasets, where we achieve state-of-the-art results. With minimal modifications, our model can also be extended for video segmentation and action recognition.
@article{arxiv.2006.00830,
title = {Temporal Aggregate Representations for Long-Range Video Understanding},
author = {Fadime Sener and Dipika Singhania and Angela Yao},
journal= {arXiv preprint arXiv:2006.00830},
year = {2020}
}