Recognising action as clouds of space-time interest points

By Matteo Bregonzio, Shaogang Gong, Tao A. Xiang. 2009 IEEE Conference on Computer Vision and Pattern Recognition, 2009

423 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Control and Systems Engineering

View Paper: https://www.semanticscholar.org/paper/09cb28e2cb1b63b78029c724e86ceb797d773b9e

How researchers see this paper

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

In Computer Vision and Pattern Recognition

In Control and Systems Engineering

Citations in the corpus

Action recognition by dense trajectories (2011, background): “Experimental results show a significant improvement over the state of the art on four datasets of varying difficulty, i.e. KTH, YouTube, Hollywood2 and UCF sports.”
Human Action Recognition From Various Data Modalities: A Review (2020, neutral): “was introduced in [287] for action classification.”
Fast Exact Hyper-graph Matching with Dynamic Programming for Spatio-temporal Data (2014, positive): “We therefore used 2D Gabor filters [9] and extracted at least 20 points per frame.”
Progressive Teacher-Student Learning for Early Action Prediction (2019, background): “Some representative handcrafted features like Cuboids [7] [44] [6], interest point clouds [4], 3DHOG [20], SIFT [33], and dense trajectory [41] etc.”
Human activity analysis (2011, background): “2009; Ryoo and Aggarwal 2009b] and the Weizmann dataset [Blank et al. 2005; Niebles et al. 2006; Scovanner et al. 2007; Rodriguez et al. 2008; Bregonzio et al. 2009].”
Real-time Action Recognition by Spatiotemporal Semantic and Structural Forests (2010, background): “17 sequence Point clouds (2009) [11] 95.”
Evaluation of Local Spatio-temporal Features for Action Recognition (2009, neutral): “[4] obtain 93.2% with a 2D Gabor ﬁlter based detector.”
Inverse Dynamics for Action Recognition (2013, methodology): “We chose two other methods [53], [54] as additional baseline methods for comparison.”
A survey of video datasets for human action and activity recognition (2013, background): “Finally, Section 5 presents the conclusion of this survey.”
Slow Feature Analysis for Human Action Recognition (2012, methodology): “In this paper, we introduce the SFA framework to the problem of human action recognition by incorporating the discriminative information with SFA learning and considering the spatial relationship of body parts.”
Video-Based Human Behavior Understanding: A Survey (2013, methodology): “to increase the robustness of the method [83], [90].”
Body Surface Context: A New Robust Feature for Action Recognition From Depth Videos (2014, background): “existing works [1], [2], [17]–[20] have achieved satisfactory”
Dense Trajectories and Motion Boundary Descriptors for Action Recognition (2013, positive): “Bregonzio et al (2009) have extended this approach with 2D Gabor ﬁlters of different orientations.”
A survey on vision-based human action recognition (2010, background): “[13] first calculate the difference between subsequent frames to estimate the focus of”
Person Re-identification by Video Ranking (2014, background): “One common representation is constructed based on space-time interest points [26,10,46,5].”
Learning a hierarchy of discriminative space-time neighborhood features for human action recognition (2010, methodology): “Almost all reported numbers use this setup, though [9] also report leave-one-out accuracy, while [3, 5, 21] exclusively use leave-one-out.”
Static and space-time visual saliency detection by self-resemblance. (2009, methodology): “Using a feature matrix consi sti g of a set of feature vectors provides more discriminative power than using a single feature vecto r as also pointed out in [41], [3].”
RESOUND: Towards Action Recognition Without Representation Bias (2018, background): “Early datasets (Weizmann [2], KTH [14]) were collected in controlled environments, minimizing static biases.”