Not all video frames are equally informative for recognizing an action. It is computationally infeasible to train deep networks on all video frames when actions develop over hundreds of …
H Cao, S Yu, J Feng - arXiv preprint arXiv:1910.02533, 2019 - arxiv.org
Although CNN has reached satisfactory performance in image-related tasks, using CNN to process videos is much more challenging due to the enormous size of raw video streams. In …
Most existing action recognition models are large convolutional neural networks that work only with raw RGB frames as input. However, practical applications require lightweight …
Motion has shown to be useful for video understanding, where motion is typically represented by optical flow. However, computing flow from video frames is very …
H Hu, W Zhou, X Li, N Yan, H Li - ACM Transactions on Multimedia …, 2020 - dl.acm.org
In video action recognition, motion is a very crucial clue, which is usually represented by optical flow. However, optical flow is computationally expensive to obtain, which becomes …
H Cheng, Y Guo, L Nie, Z Cheng… - Proceedings of the 31st …, 2023 - dl.acm.org
Training an effective video action recognition model poses significant computational challenges, particularly under limited resource budgets. Current methods primarily aim to …
We present SlowFast networks for video recognition. Our model involves (i) a Slow pathway, operating at low frame rate, to capture spatial semantics, and (ii) a Fast pathway, operating …
W Wu, D He, X Tan, S Chen… - Proceedings of the …, 2020 - openaccess.thecvf.com
Though action recognition in videos has achieved great success recently, it remains a challenging task due to the massive computational cost. Designing lightweight networks is a …
Y Zhi, Z Tong, L Wang, G Wu - Proceedings of the IEEE/CVF …, 2021 - openaccess.thecvf.com
Frame sampling is a fundamental problem in video action recognition due to the essential redundancy in time and limited computation resources. The existing sampling strategy often …