paper-with-me

Papers

Motion Representation Using Residual Frames with 3D CNN

2020-06-21 · Li Tao, Xueting Wang, Toshihiko Yamasaki

Recently, 3D convolutional networks (3D ConvNets) yield good performance in action recognition. However, optical flow stream is still needed to ensure better performance, the cost of which is very high. In this paper, we propose a fast but effective way to extract motion features from videos utilizing residual frames as the input data in 3D ConvNets. By replacing traditional stacked RGB frames with residual ones, 35.6% and 26.6% points improvements over top-1 accuracy can be obtained on the UCF101 and HMDB51 datasets when ResNet-18 models are trained from scratch. And we achieved the state-of-the-art results in this training mode. Analysis shows that better motion features can be extracted using residual frames compared to RGB counterpart. By combining with a simple appearance path, our proposal can be even better than some methods using optical flow streams.

📄 PDF Abstract BibTeX arXiv:2006.13017

Code (3)

BestJuly/IIC pytorch
BestJuly/Inter-intra-video-contrastive-learning pytorch
BestJuly/VCP pytorch

Tasks

Action RecognitionOptical Flow Estimation

Similar Papers 제목 키워드 기반

Residual Frames with Efficient Pseudo-3D CNN for Human Action Recognition

2020-08-03 · Jiawei Chen, Jenson Hsiao, Chiu Man Ho

Human action recognition is regarded as a key cornerstone in domains such as surveillance or video understanding. Despite recent progress in the development of end-to-end solutions for video-based action recognition, ach…

Action RecognitionOptical Flow EstimationTemporal Action LocalizationVideo Understanding

Rethinking Motion Representation: Residual Frames with 3D ConvNets for Better Action Recognition

2020-01-16 · Li Tao, Xueting Wang, Toshihiko Yamasaki

Recently, 3D convolutional networks yield good performance in action recognition. However, optical flow stream is still needed to ensure better performance, the cost of which is very high. In this paper, we propose a fas…

Action RecognitionOptical Flow Estimation

Generative Motion Infilling From Imprecisely Timed Keyframes

2025-03-02 · Purvi Goel, Haotian Zhang, C. Karen Liu, Kayvon Fatahalian

Keyframes are a standard representation for kinematic motion specification. Recent learned motion-inbetweening methods use keyframes as a way to control generative motion models, and are trained to generate life-like mot…

End-to-end Optimized Video Compression with MV-Residual Prediction

2020-05-26 · XiangJi Wu, Ziwen Zhang, Jie Feng, Lei Zhou 외

We present an end-to-end trainable framework for P-frame compression in this paper. A joint motion vector (MV) and residual prediction network MV-Residual is designed to extract the ensembled features of motion represent…

MS-SSIMSSIMVideo Compression

End-to-End Compressed Video Representation Learning for Generic Event Boundary Detection

2022-03-29 · CVPR 2022 1 · CongCong Li, Xinyao Wang, Longyin Wen, Dexiang Hong 외

Generic event boundary detection aims to localize the generic, taxonomy-free event boundaries that segment videos into chunks. Existing methods typically require video frames to be decoded before feeding into the network…

Boundary DetectionGeneric Event Boundary DetectionRepresentation Learning