Long-Term Feature Banks for Detailed Video Understanding

Chao-Yuan Wu, Christoph Feichtenhofer, Haoqi Fan, Kaiming He, Philipp Krähenbühl +1 more
12/12/2018
cs.CV

Official code: facebookresearch/video-long-term-feature-banks (384 stars, Apache-2.0, last commit 2021-08-30).

Abstract

To understand the world, we humans constantly need to relate the present to the past, and put events in context. In this paper, we enable existing video models to do the same. We propose a long-term feature bank---supportive information extracted over the entire span of a video---to augment state-of-the-art video models that otherwise would only view short clips of 2-5 seconds. Our experiments demonstrate that augmenting 3D convolutional networks with a long-term feature bank yields state-of-the-art results on three challenging video datasets: AVA, EPIC-Kitchens, and Charades.

View on arXivView PDF

Code Implementations(1)

facebookresearch/video-long-term-feature-banksOfficial100%
Apache-2.0

Long-Term Feature Banks for Detailed Video Understanding

38462C++, CudaMar 6, 20195 years agoApache-2.0

Similar Papers

MA-LMM: Memory-Augmented Large Multimodal Model for Long-Term Video Understanding

87% match1 repo3532024

VideoMamba: State Space Model for Efficient Video Understanding

86% match1 repo1.1k2024

Precise Event Spotting in Sports Videos: Solving Long-Range Dependency and Class Imbalance

85% match2025

LongVLM: Efficient Long Video Understanding via Large Language Models

84% match1 repo1082024

HAT: History-Augmented Anchor Transformer for Online Temporal Action Localization

84% match1 repo202024

VideoMem: Enhancing Ultra-Long Video Understanding via Adaptive Memory Management

83% match2025

Cite this paper

@article{wu2018longterm,
  title  = {Long-Term Feature Banks for Detailed Video Understanding},
  author = {Chao-Yuan Wu and Christoph Feichtenhofer and Haoqi Fan and Kaiming He and Philipp Krähenbühl and Ross Girshick},
  year   = {2018},
  eprint = {1812.05038},
  archivePrefix = {arXiv},
  url    = {https://arxiv.org/abs/1812.05038}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/1812.05038.svg)](https://codeofpaper.com/paper/1812.05038)

Discussion