Z
Zheng Shou
Researcher at Columbia University
Publications - 36
Citations - 3652
Zheng Shou is an academic researcher from Columbia University. The author has contributed to research in topics: Action (philosophy) & Computer science. The author has an hindex of 16, co-authored 34 publications receiving 2617 citations. Previous affiliations of Zheng Shou include Facebook & Wuhan University.
Papers
More filters
Proceedings ArticleDOI
Temporal Action Localization in Untrimmed Videos via Multi-stage CNNs
TL;DR: Wang et al. as mentioned in this paper exploit the effectiveness of deep networks in temporal action localization via three segment-based 3D ConvNets: a proposal network identifies candidate segments in a long video that may contain actions, a classification network learns one-vs-all action classification model to serve as initialization for the localization network, and a localization network fine-tunes the learned classification network to localize each action instance.
Proceedings ArticleDOI
CDC: Convolutional-De-Convolutional Networks for Precise Temporal Action Localization in Untrimmed Videos
TL;DR: In this paper, a Convolutional-De-Convolutional (CDC) network is proposed for temporal action localization, which performs spatial upsampling and spatial downsampling operations simultaneously to predict actions at the frame-level granularity.
Posted Content
CDC: Convolutional-De-Convolutional Networks for Precise Temporal Action Localization in Untrimmed Videos
TL;DR: A novel Convolutional-De-Convolutional (CDC) network that places CDC filters on top of 3D ConvNets, which have been shown to be effective for abstracting action semantics but reduce the temporal length of the input data.
Proceedings ArticleDOI
Single Shot Temporal Action Detection
Tianwei Lin,Xu Zhao,Zheng Shou +2 more
TL;DR: Wang et al. as discussed by the authors proposed a Single Shot Action Detector (SSAD) network based on 1D temporal convolutional layers to skip the proposal generation step via directly detecting action instances in untrimmed video.
Posted Content
ConvNet Architecture Search for Spatiotemporal Feature Learning.
TL;DR: This paper presents an empirical ConvNet architecture search for spatiotemporal feature learning, culminating in a deep 3D Residual ConvNet that outperforms C3D by a good margin on Sports-1M, UCF101, HMDB51, THUMOS14, and ASLAN while being 2 times faster at inference time, 2 times smaller in model size, and having a more compact representation.