mlsanityData to 7 Oct 2026

Delta Sampling R-BERT for limited data and low-light action recognition

arXiv 2107.05202 ↗First posted 12 Jul 2021cs.CVcited by 7

Not on the map. Why?

Abstract (arXiv)

We present an approach to perform supervised action recognition in the dark. In this work, we present our results on the ARID dataset. Most previous works only evaluate performance on large, well illuminated datasets like Kinetics and HMDB51. We demonstrate that our work is able to achieve a very low error rate while being trained on a much smaller dataset of dark videos. We also explore a variety of training and inference strategies including domain transfer methodologies and also propose a simple but useful frame selection strategy. Our empirical results demonstrate that we beat previously published baseline models by 11%.

Key facts. Cited by 7 papers in the map (data through 7 Oct 2026). Builds on 38 of its 69 references that are in the map. First posted on arXiv 12 Jul 2021.

Builds on

38 of 69 references.

References · 38

  1. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,204
  2. 2019PyTorch: An Imperative Style, High-Performance Deep Learning Librarycited by 20,160
  3. 2019EfficientNet: Rethinking Model Scaling for Convolutional Neural Networkscited by 6,415
  4. 2017Squeeze-and-Excitation Networkscited by 4,763
  5. 2017Quo Vadis, Action Recognition? A New Model and the Kinetics Datasetcited by 3,259
  6. 2017The Kinetics Human Action Video Datasetcited by 1,894
  7. 2018SlowFast Networks for Video Recognitioncited by 1,862
  8. 2021Is Space-Time Attention All You Need for Video Understanding?cited by 1,609
  9. 2017A Closer Look at Spatiotemporal Convolutions for Action Recognitioncited by 1,139
  10. 2017The "something something" video database for learning and evaluating visual common sensecited by 1,092
  11. 2017The 2017 DAVIS Challenge on Video Object Segmentationcited by 942
  12. 2018Scaling Egocentric Vision: The EPIC-KITCHENS Datasetcited by 789
26 more
  1. 2019On the Variance of the Adaptive Learning Rate and Beyondcited by 763
  2. 2015Cyclical Learning Rates for Training Neural Networkscited by 627
  3. 2017Can Spatiotemporal 3D CNNs Retrace the History of 2D CNNs and ImageNet?cited by 551
  4. 2020Zero-Reference Deep Curve Estimation for Low-Light Image Enhancementcited by 541
  5. 2016Hollywood in Homes: Crowdsourcing Data Collection for Activity Understandingcited by 535
  6. 2017Rethinking Spatiotemporal Feature Learning: Speed-Accuracy Trade-offs in Video Classificationcited by 521
  7. 2019EnlightenGAN: Deep Light Enhancement without Paired Supervisioncited by 466
  8. 2018Learning to See in the Darkcited by 459
  9. 2017AVA: A Video Dataset of Spatio-temporally Localized Atomic Visual Actionscited by 449
  10. 2019Lookahead Optimizer: k steps forward, 1 step backcited by 333
  11. 2019Video Classification with Channel-Separated Convolutional Networkscited by 290
  12. 2015Beyond Short Snippets: Deep Networks for Video Classificationcited by 233
  13. 2015LLNet: A Deep Autoencoder Approach to Natural Low-light Image Enhancementcited by 224
  14. 2018A disciplined approach to neural network hyper-parameters: Part 1 -- learning rate, batch size, momentum, and weight decaycited by 223
  15. 2017Deep Bilateral Learning for Real-Time Image Enhancementcited by 200
  16. 2021Revisiting ResNets: Improved Training and Scaling Strategiescited by 158
  17. 2019Large-scale weakly-supervised pre-training for video action recognitioncited by 130
  18. 2020Gradient Centralization: A New Optimization Technique for Deep Neural Networkscited by 82
  19. 2017ActionVLAD: Learning spatio-temporal aggregation for action classificationcited by 73
  20. 2019Learning Spatio-Temporal Representation with Local and Global Diffusioncited by 61
  21. 2016VideoLSTM Convolves, Attends and Flows for Action Recognitioncited by 60
  22. 2018Distort-and-Recover: Color Enhancement using Deep Reinforcement Learningcited by 52
  23. 2020ARID: A New Dataset for Recognizing Action in the Darkcited by 42
  24. 2017Aesthetic-Driven Image Enhancement by Adversarial Learningcited by 38
  25. 2017WESPE: Weakly Supervised Photo Enhancer for Digital Camerascited by 32
  26. 2018Image Correction via Deep Reciprocating HDR Transformationcited by 30

Built on by

Cited by year

Citing papers per year: 2022: 1; 2023: 5; 2024: 0; 2025: 1.

7 papers.

  1. 2023Division Gets Better: Learning Brightness-Aware and Detail-Sensitive Representations for Low-Light Image Enhancementcited by 4
  2. 2022Going Deeper into Recognizing Actions in Dark Environments: A Comprehensive Benchmark Studycited by 3
  3. 2023Bio-Inspired Simple Neural Network for Low-Light Image Restoration: A Minimalist Approachcited by 0
  4. 2023Dual Degradation-Inspired Deep Unfolding Network for Low-Light Image Enhancementcited by 0
  5. 2023IndGIC: Supervised Action Recognition under Low Illuminationcited by 0
  6. 2023Diving into Darkness: A Dual-Modulated Framework for High-Fidelity Super-Resolution in Ultra-Dark Environmentscited by 0
  7. 2025Seeing in the Dark: A Teacher-Student Framework for Dark Video Action Recognition via Knowledge Distillation and Contrastive Learningcited by 0