Temporal Action Localization in Untrimmed Videos via Multi-stage CNNs

By Zheng Shou, Dongang Wang, Shih-Fu Chang. 2016

973 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Endocrinology, Diabetes and Metabolism, Model Architectures

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

In Endocrinology, Diabetes and Metabolism

In Model Architectures

Citations in the corpus

Towards Automatic Learning of Procedures From Web Instructional Videos (2017, methodology): “In terms of proposal selection, a commonly adopted method in object detection (Ren et al. 2015) or action detection (Shou, Wang, and Chang 2016) is NMS.”
End-to-End Dense Video Captioning with Masked Transformer (2018, neutral): “[27] propose and classify proposal candidates directly over video frames in a sliding window fashion, which is computationally expensive.”
Skeleton-Based Online Action Prediction Using Scale Selection Network (2019, background): “[70] adopted multi-scale windows for action detection”
RGB-D-based Human Motion Recognition with Deep Learning: A Survey (2017, methodology): “Thus, proposal-based methods (Shou et al., 2016; Wang et al., 2017a) can be adopted for offline processing.”
A Closer Look at Spatiotemporal Convolutions for Action Recognition (2017, background): “3D CNNs features were also demonstrated to generalize well to other tasks, including action detection [28], video captioning [24], and hand gesture detection [23].”
TALL: Temporal Activity Localization via Language Query (2017, methodology): “Temporal action boundary regression As described before, we implemented a temporal localization loss function similar to the one in SCNN [26], which consider clip overlap.”
Asynchronous Temporal Fields for Action Recognition (2016, background): “With the advent of deep learning, learning representations from data has been extensively studied [14, 15, 44, 57, 52, 53, 24, 7, 61, 55, 40, 3].”
Motion-Appearance Co-memory Networks for Video Question Answering (2018, methodology): “[29] presented a multi-stage Segment-CNN model to generate action proposals and localize actions in videos.”
Temporal Context Network for Activity Localization in Videos (2017, methodology): “Here, we compare our temporal context network with DAPs[6], PSDF[15], FG[27] SCNN[30] and CDC[28].”
Attentive Moment Retrieval in Videos (2018, background): “Shou et al. [28] proposed an end-to-end segment-based 3D Convolutional Neural Network (CNN) framework, which outperforms other Recurrent Neural Network (RNN)-based methods by capturing spatio-temporal information simultaneously.”
SSNet: Scale Selection Network for Online 3D Action Prediction (2018, methodology): “Also, instead of using one fixed scale [77] or multi-scale multi-pass scans [55, 84], we propose a novel SSNet for online prediction, which is supervised to choose the proper window for prediction at each temporal step.”
ActionSpotter: Deep Reinforcement Learning Framework for Temporal Action Spotting in Videos (2020, methodology): “between detection and spotting: while CDC [32] is more efficient in detection than M-CNN[31], it is the opposite for spotting.”
Temporal Action Detection with Structured Segment Networks (2017, methodology): “2015), SCNN-prop (Shou et al. 2016), DAP (Escorcia et al.”
Graph Distillation for Action Detection with Privileged Modalities (2017, methodology): “gress in action classiﬁcation, recent work has tackled the temporal action detection task. Previous work includes frame-level or sliding window classiﬁcation methods [14,19,36], proposalbased methods [11,49,69], and end-to-end single-pass frameworks [63,2]. The success in RGB video understanding has”
BSN: Boundary Sensitive Network for Temporal Action Proposal Generation (2018, methodology): “Table 6: Action detection results on testing set of THUMOS14 in terms of mAP@tIoU , where classification results generated by UntrimmedNet [43] and SCNN-classifier [32] are combined with proposals generated by BSN and other methods.”
SCSampler: Sampling Salient Clips From Video for Efficient Action Recognition (2019, background): “More closely related to our task is the problem of action localization [22, 39, 38, 54, 61], where the objective is to localize the temporal start and end of each action within a given untrimmed video and to recognize the action class.”
Rethinking the Faster R-CNN Architecture for Temporal Action Localization (2018, methodology): “Inspired by the recent success of region-based detectors in object detection [18, 17], many recent approaches adopt a two-stage, proposal-plus-classification framework [6, 36, 12, 3, 4, 35, 56], i.”